Around the mid-90s, the Republic's annual investment in the telecommunications industry claimed to have exceeded its annual defense spending, and if we delved into the root causes, a large proportion of them were made use of interest-free and low-interest long-term loans from overseas.

For example, in the three years from 91 to 94, the Republic of China used more than $2.5 billion in overseas loans on a single device of digital switches, and introduced more than 10 million lines of international brands of switches.

In addition to imported switches, in the mid-to-early nineties, the only company in the Republic that could produce digital switches was a joint venture company called Wu Songbel.

Without Liang Yuan, this joint venture company would have grown into a monopoly giant in the Republic's telecommunications market in the 90s, and at its peak, Bell's products once accounted for 60 percent of the Republic's program-controlled telephone network equipment.

With the success of the supercomputing project, the benefits of large-scale cutting-edge technology from the basic industry have begun to be reflected, such as the current Inspur 40 million times per second, with an external quotation of 2.4 million units, which can be regarded as a high-performance workstation of the Inspur R&D subsidiary achievements.

No matter how expensive a simple supercomputer is, three or four hundred million units will basically reach the sky, and if you talk about the volume, it is destined to be inferior to the Inspur workstation.

Limited to the elite version of the C3 processor high-frequency chip yield is not high, the production capacity of Yuanjia supercomputer is expected to be less than ten per year, even if the 2.0 micron process technology that has been involved with Motorola and Siemens for a year will be approved for introduction immediately, and the formal production of the equipment will be fast for eight months after installation and commissioning, and the yield rate of the new line will reach an acceptable commercial level, which is already twelve or even fourteen months later.

At present, the supercomputer with a wave computing speed of 3 billion times per second can barely be regarded as touching the tip of the tail of the international mainstream level for the time being, but in terms of the development speed of the chip, a year later, that is, in the middle of 91, if the supercomputer of the wave wants to remain competitive, it must at least increase the computing speed to more than 6 billion times, and it must break through the 10 billion mark in 92, which is almost an endless marathon, and a little relief means falling behind.

The development of the next-generation processor C4 has been on the right track, thanks to the accumulation of technology from Fang Xingdong and the introduction of MIPS in the early stage, and the estimated development time of the new generation processor is as long as 20 months.

If the C4 processor is successfully developed, the supercomputing system that has been outdated at that time will definitely face an upgrade, fortunately, the supercomputing architecture of the inspur is advanced, but the ability of the supercomputer to process data after the upgrade has increased greatly, and the router must also be upgraded, from the original router elite version of the C3 main control chip to the elite version of the C4 main control chip, which is equivalent to falling into a new round of cycle, the output of the top supercomputer still can't be raised, it is still ten per year.

To solve this bottleneck, unless the 1.0 micron advanced process is directly introduced, unfortunately, the possibility of directly introducing overseas top process technology is not as good as the possibility of independent breakthroughs in Liang Tuhao's Krypton Jinmeng's domestic technology.

Of course, it's just that the top batch of supercomputing is subject to output, and the output of other similar high-performance workstations is large, but limited to the Hong Kong-based current collector 5.0 micron has lagged behind the international mainstream production line process technology, the power consumption and heat generation of the Inspur high-performance workstation are extremely large, basically an ordinary studio, put two workstations on the winter, and do not need heating.

But even so, the Inspur workstation will also bring great benefits to the Gangji power collector, the quotation range from more than 20 million to more than 50 million, by eight to nineteen Inspur high-performance workstations composed of three to nine billion times per second of small computing capacity, has been enough to meet the needs of more than ten industries such as finance, taxation, stock market, scientific research, geology, telecommunications and so on for high-performance computers.

In terms of hardware performance, Inspur's current small technology cluster is basically equivalent to the low-end X-series or I-series products in IBM's world-renowned high-performance mainframes in the future, such as fully meeting the data processing needs of provincial and municipal banks.

Of course, for the data processing of the backbone nodes of national finance, energy, telecommunications, etc., only supercomputing can be done in the product line of Inspur, and the software has to continue to work hard, fortunately, the current time has only entered the nineties, and IBM's commercial mainframe has just started, leaving enough time for Yuanjia.

On May 18 last month, after Inspur announced the successful development of the Republic's supercomputer of 3 billion times per second, less than 48 hours later, Wu Zhonghua, on behalf of the Chinese Academy of Sciences, went directly to Liang Yuan's head.

Wu Zhonghua, who is well aware of Yuanjia's internal situation, has long learned from Li Yuanling that the annual output of supercomputing is not high due to the fact that the large-traffic router is too difficult to deliver.

The first supercomputer has been pulled away by the military's aircraft, and the second supercomputer is scheduled by Brilliance, the priority of the two must be lightning and even Liang Yuan does not dare to express any opinions, but it is worth discussing who the third supercomputer will give to.

When people mention the republic in the 90s, they will use adjectives that may include many beautiful words such as simplicity, vigor, and diligence, but poverty and backwardness are certainly one of the top words, but it is difficult for anyone to believe that the republic is actually very fashionable in the pursuit of the application of advanced technology, but it is limited to national strength and many beautiful ideas cannot be supported.

For example, in the mid-80s, the future iron-faced prime minister of the republic proposed the possibility of a fully electronic government office, but at that time the republic had neither the technology nor the capital, and there were countless more important directions than electronic office waiting to be invested in compared to the foreign exchange funds that could be used.

Old Man Wu really came right, just found Liang Yuan and didn't say a word, Su Liangyu's phone chased directly into Liang Yuan's study.

The Ministry of Finance and the central bank are very interested in the supercomputer of the wave, and the central bank did not even ask the price of the supercomputer, and crossed the relationship between the tide and the trust in the front desk to find Su Liangyu's head, and directly said that he could book two units in full.

No matter what year the officials of the Republic are, there are naturally those who work hard to be promoted, and it is obvious that the matter of supercomputing has not only been concerned by the military of the Republic since the beginning of research and development, but it is limited to the military's secrecy system, and countless other yamen just know that there is a great chance of success of this project, and other details will definitely not be known.

But with the announcement of the news by the wave, prepared smart people will naturally try to seize the opportunity in advance, not to mention that under someone's push, the Republic has more national-level data divisions than in the original history, and the current functions and powers of the data division With the development of the torch plan, there is already a trend of moxibustion in the field of science and technology in the Republic.

In the early 90s, almost all of the people who could really understand what supercomputers could do were the top elites in various industries, and these people were definitely more familiar with the application of supercomputers in this industry than Liang Yuan, after all, they were professionals.

There are not many departments in the republic that are qualified to use large sums of foreign exchange, but the yamen that is qualified to use large sums of renminbi has gone to the sea, and the data centers of the head offices of several major banks in China alone have to be built with computers of this level.

The technical details of the research and development process of supercomputing are confidential, but the output of supercomputing is not high, and it is definitely not within the scope of confidentiality.

Compared with the wealthy Ministry of Finance and the more wealthy central bank, the Chinese Academy of Sciences must be a poor ghost, but it is rare for the old man to ask for someone for the first time, and Liang Yuan can't refuse anything.

Originally, Liang Yuan had already planned the location of the third supercomputer, but Liang Yuan didn't expect that the supercomputer that had just gone offline could become like this.

Liang Yuan has lived in the information age for a long time, and looking at the humble and crude information exchange and communication methods in the early nineties, he always has the illusion of looking at the prehistoric era, and naturally ignores that some top industry elites can really see the future of the industry in ten or twenty years.

In Liang Yuan's plan, Liang Yuan, the third supercomputer, plans to sell it half and half give it away to the National Meteorological Administration in Beijing.

The flood is coming, the flood is coming again, the flood is coming again~.

Although Liang Yuan has not personally experienced what kind of mood it will be when the yellow flood flows in front of his house in the north of the Republic, since he was a teenager, he has seen similar hardships repeated on TV almost every summer, which is enough to move Liang Yuan.

When there was no supercalculation, even if Liang Yuan had the intention to prompt something, in the end he still didn't dare to say anything, didn't dare to do anything, and the logical dead knot tied Liang Yuan's hands and feet fiercely.

Now, with the supercomputer offline, Liang Yuan can finally be successfully liberated.

In the future meteorological community, there is an assumption that if humans have the technology to monitor every gas molecule on the earth, then with sufficient computer power, human beings can accurately predict climate change at every point in the world every day before a sudden change such as a large-scale volcanic eruption or asteroid impact occurs.

The performance of supercomputers has always been closely related to a country's ability to predict the weather, and through supercomputing, Yuanjia, which has strong economic strength, is fully capable of giving early warning of future disasters in the name of science.

For example, the Huai River flood in 91 years is close at hand.

At that time, not everyone was unaware of the Huai River flood, and a meteorological engineer from the Central Meteorological Bureau of the Republic of China found anomalies by comparing historical meteorological records with observing early Pacific satellite cloud images.

However, in the face of the early mobilization of nearly 100 million people involving hundreds of thousands of square kilometers, who will bear the economic losses once the economic loss is wrongly judged?

This kind of economic loss does not refer to the shutdown of the factory or anything, even if each person prepares a paltry five packs of Huafeng instant noodles in advance, the cost will be 300 million, and how can the preparation of five packs of instant noodles in the face of a large water that lasts for more than a month.

After several rounds of heated debates among experts in related fields, the meteorological engineer who first discovered the clues chose to give up his own views and handed over the final choice to God.

There is nothing to blame for the last choice of this meteorological engineer, although human forecasting in the field of meteorology is far stronger than in the field of earthquakes, but that is after the great improvement of computer performance in the new century, and it is limited to the small area of a certain city, and the accurate prediction of the long-term trend of a large area is still a world-class problem.

Technically, it is easy to understand that the larger the area, the more gas molecules need to be monitored, and the more computing power is required.

For the Central Meteorological Bureau, which can be said to be at the forefront of human scientific research since its establishment, it can almost be regarded as a research unit directly under the Chinese Academy of Sciences.

Liang Yuan did not hide his intentions, and directly told Wu Zhonghua about his original plan, and finally after the Central Weather Bureau must obtain at least 25% of the supercomputing time, Liang Yuan sold the third supercomputer produced by Inspur to the Chinese Academy of Sciences at a seven-percent discount.

Then, someone was completely wilted in the face of the almost surging tide of supercomputer rush purchases, and after throwing the pot to Su Liangyu, Liang Yuan didn't even go home and went directly to the operation room prepared by He Yunwei for the three of them in the Shangri-La Hotel.

Due to the high correlation of these two technologies, Inspur has mastered the difficult router, and it is natural to stand at the peak of the research and development of the switch, and even because of the high starting point, the Inspur is like a bamboo in the research and development of the switch, and there is almost no lag directly breaking through the current international mainstream 100,000 line technology mark.

To a certain extent, Liang Yuan's command of the Hong Kong-based power collection and Inspur supercomputer projects is similar to the United States landing on the moon, although the cost is huge and the risk is high, but once it succeeds, the harvest is also extremely rich, just like Inspur through the research and development of top supercomputer projects, almost overnight became a world-class telecommunications equipment manufacturer.

"In addition to feeding Jiajia and Fei Fei these days to prevent the two of them from becoming eggs, I also have to prepare for going to Wusong in July to participate in the finals of the Youth Science Competition, I didn't look at the recent experimental data of the program-controlled switch, the test of the group's products has always had to go through your side, Comrade Xiao Nie told me the truth about the switch test, I won't tell Wang Xin and Liu Feiyang if there is anything bad." Liang Yuan said with a smile.

"The progress of routers and supercomputing is big or small, but I have time to ask, the progress of the switch in the past two years, I really haven't heard the big or young ask, and I think that the big and young don't look down on the small market of program-controlled switches. ”

Nie Xiaotian sighed, and then continued: "The 10,000-line switch will be left on my side for testing, at present, in terms of the use effect within the group, it is great, whether it is sound quality or connection rate or drop rate is very good, there is also a top product of 100,000 lines that was sent to the city post office for testing, the equipment itself has not had any problems, but because the city's switch brand has three countries and four standards, the phone drop rate is very high, and the voice distortion problem is also very prominent." ”

"If you want me to say big or little, unless the city can replace other brands of switches with our brand, the wave side will have to continue to strengthen the compatibility of equipment. ”