Chapter 167: The Curtain Opens (16)
If there is still a hierarchy in Yuanjia’s conference room, the Mike in Li Guowei’s hand is the most obvious example.
Only the bosses at the Yuanjia Conference are qualified to pick up the microphone and explain their views to everyone in the room. Most of the nearly 100 senior executives who participated in the Yuanjia Conference this time were listeners.
Because the supercomputer is about to be offline, although Wang Xin and Liu Feiyang, who are currently in charge of Inspur's business, have received generous year-end rewards, they are still attending the meeting and still have no say.
If we calculate based on the comprehensive calculation of scale, assets, brand and other factors in the industry, we can participate in the management backbone of the Yuanjia Conference this time, which one is not the only one? For example, the Mandarin Oriental Hotel, which is responsible for Song Xiaowei, can be ranked in the whole of Asia. However, no matter how well the Mandarin Oriental Hotel develops or goes bankrupt, it will not affect the core operations of Yuanjia Enterprise.
Therefore, attendance is an attendance. Only Yuanjia’s absolute core department is qualified to concentrate resources and to customize Yuanjia’s future.
Of course, if there are any confusion or suggestions during the meeting, many auditors can communicate in advance within the small circle of the region, and then explain it by the president of the conference level in the region.
More importantly, the many attendees, that is, the CEO in the future Yuanjia reserve, is also a catfish created by Liang Yuan for members of the Yuanjia Conference. The existence of these reserve forces reminds that the presidents of the conference have long been successful, and they are not good at doing well, but they are everywhere. They can easily pull up without any transition, and they can directly take over.
Listening to Li Guowei briefly introduce the ins and outs of the Yuanjia Conference-level boss, Ou Ruiyang smacked his lips and asked with some emotion: "It turns out that the difference between the boss and the boss is quite big. Guowei, do you come to COMAC in advance before I get the right to speak now?"
"Originally, there were only two people who received this treatment from COMAC. One was the chairman Liang Haiping, the other was COMAC president Zhou Yuanhang. However, my work relationship was transferred to the gas engine laboratory by the boss's wife last month. Now I can exercise the authority of the president of the conference under the name of FAW."
"Isn't Guowei responsible for the subsonic aerodynamics of aircraft engines at COMAC? Why did he change his career to build turbines on the ground?"
After hearing Li Guowei say this, Ou Ruiyang was very surprised. Even the information about Li Guowei's rapid promotion in Yuanjia system was ignored.
The technical foundation of FAW is the Rolls-Royce Aircraft Engine of the UK. Although FAW's business scope is currently limited to ground power generation, transportation and other fields, Li Yuanling's blue sky sentiment has never changed.
After the Nansha battle, the various things in this war were not diluted, but instead caused more and more trouble in the republican world.
Modern wars can still be fought like this???
After entering the city, Tujiao Road found that there were too many things to buy. What was even more detrimental is that the people in the city were not particular and always discriminated against Tujiao Road's shopping environment, which was very unfriendly. In addition, Tujiao Road had no money in its pocket, so the banner of self-reliance naturally became the new favorite of Tujiao Road.
In the self-reliance project, the Republic Navy was still gritting its teeth to develop the J-H-7 based on the Scot Aircraft Aircraft engine. The Republic Air Force made a comeback. The H-6 modification that could fly to Guam a few years ago was found by the Air Force from the old paper pile.
The H-6 is similar to the J-7 fighter. Its technical level has been completely understood by the Republic. The only bottleneck that restricts West Air Flight from improving in a variety of ways is that the Republic does not have a large duct turbofan engine that can work stably.
In the past six months, I don’t know where Ning Lei got a lot of technical information about the original version of the SBE Aviation engine, namely the civilian version of the Dahando SBE Aviation engine, and divided it into several batches and gave it to Li Yuanling.
Although Ning Lei didn't say a word about developing the New Aviation Engine, Sima Zhao's intention is well known to everyone.
After Liang Yuan bought Yihe, Li Yuanling's enthusiasm suddenly improved a lot. Although the Republican Fang Wei was a little poor, for Yuanjia, who has a lot of infrastructure, it is also an acceptable option to pay with labor in addition to cash.
The main forces of the construction of Changbai Island factory are the world of engineers in addition to China Construction Third Bureau. At this time, the tide of millions of disarmament in the Republic had just receded. For the relevant military leaders, if they have work to prove their value, they can keep the establishment unchanged. Therefore, it is also optimistic that the infrastructure teams under the military will participate in this project.
Ning Lei took the time to attend the Yuanjia Conference under the guise of celebrating the New Year in advance, because the supercomputing project was implemented very smoothly.
At this time, most of the existing supercomputers internationally existing structures were multi-vector flow megaphones, such as the American Clay computer, ibm computer, Japanese nec computer, Republic Galaxy computer, etc., which are typical representatives of giant machines.
However, due to the high development costs of giant machines and low expansion, there are three future development directions for supercomputer technology, one is multi-threading, the other is parallel processing of distributed storage, and the last is the cluster computer system used by Inspur.
At this time, it was less than eight years since the well-known American server manufacturer Sun Computer proposed that the concept of network is not only a computer, and it was still one year before the first multi-processor workstation designed with cloud concept was launched on the market. The cluster computer system was not valued by the international supercomputer community at all.
For example, in the eyes of the Blue Giant ibm, high-end, atmospheric and high-end multi-threaded giant machines are the future of supercomputers. The so-called distributed computing, in terms of the performance of personal computers in the early 1990s, is simply trying to assemble a performance sports car with bicycle parts.
At the turn of the century, the main source of the turbulence of the international information industry was that it could not find a beacon to guide the way forward, and could only rely on the inspiration of geniuses to move forward in a stumble and twists, such as the rise of Microsoft and Cisco, the collapse of Wang'an's computer, and the almost bankruptcy of ibm.
However, due to someone's bug, Yuanjia does not have any problem of making any choices on the technology development route. Even if someone is a hardware idiot, he still has an impression of the famous cloud computing.
Although Liang Yuan cannot solve any technical problems, it is enough for Yuan Jia to pretend to be a promising profession of lighthouse. As long as the direction is determined, even if the research and development progress is temporarily stuck, it will not be confused or shaken. It will only be considered by someone to be a problem of insufficient money spent.
The so-called "violent miracles" are used to break all methods, adopt a parallel computing system, a workstation based on high-performance SSC chips, and a cluster computer system formed by a high-speed Internet connection. In just one year, most of the technical nodes were overcome by the technical team led by Liu Feiyang and Wang Xin.
According to the military representatives at Inspur, the floating-point computing capability of a clustered workstation based on the c2 processor has exceeded 17 million per second, and each running workstation is composed of 86 c2 processors with a main frequency of 20mhz.
Among them, the individual workstation station is about 1.9 meters high, one meter long and 0.5 meters wide. Inspur has successively produced and assembled seven workstations and connected the seven workstations together through a high-speed fiber optic network.
After hard debugging and writing many professional software for distributed computing, the computing power of the cluster supercomputer system composed of seven base stations has exceeded the first generation of Galaxy megacomputers developed by the Republic a few years ago, with a peak computing speed of 119 million per second.
Of course, if you simply put this computing data of 119 million times per second in the international market, it may be ridiculous. Intel's 80486 chips expected to release this year has exceeded 55 million times per second. Judging from the data, just two 80486 chips are enough to match Inspur's seven Inspur workstations.
Workstations that start from the perspective of cost of use are simply inefficient, and their inputs and outputs are disproportionately disproportionately.
However, industry insiders such as Su Liangyu and Jian Zhikang were almost ecstatic about the launch of Inspur Workstation.
The main frequency of the standard version of the c2 processor is only 10mhz, and the computing speed of the processor is just over one million times per second. The computing speed of the high-main frequency c2-24mhz processor is barely close to 4 million times per second. The computing speed of the same high-main frequency 80486-100mhz processor is 55 million times per second. The computing power difference between the two is 13.75 times.
Hong Kong Basic Co., Ltd. expects that the computing power of the streamer c3-33mhz processor can reach 9 million times per second in May 1990. In the future, the main frequency of the high-end c3 reaches 66mhz, and it is optimistic that the computing speed of the high-frequency c3 processor can reach half of the 80486 computing speed.
What does this mean? The most significant feature of supercomputers using cluster computer systems is their extremely powerful expansion performance.
Under a certain upper limit, as long as the network skeleton that forms the computer cluster remains unchanged, the entire supercomputer system can completely rely on updating the computing nodes to obtain the upgrade of the supercomputer system's computing speed.
This means that the 1.1 supercomputer version composed of the same seven Inspur workstations will reach a level of 1 billion times per second, and if the future C4-level processor is selected, the computing speed of the 7 Inspur workstations can exceed 2 billion times per second.
As long as the network bandwidth problem is solved, even if you choose a C2 processor, you only need to choose the simplest brute-overlay method to push the number of operations per second of the supercomputer to more than 3 billion times. If it is upgraded to the high-main frequency C3, the simulation problem of the Republic's nuclear test will never be any more. (To be continued.)
Chapter completed!