0p0 Page 477
Page 477
……
Beijing.
Three days ago, the architecture design of the "National Computing and Information Processing System" was completed.
Tang Hua does not intend to copy OGAS...
OGAS was proposed in 1962 and rejected by Brezhnev in 1970.The reason is that it is too large and complex, and the expected effect may not be so good.
According to the concept, this thing is a network similar to or even larger than ARPANET, and it is divided into three levels of network structure. In addition to the network center of the Moscow State Planning Commission, intermediate centers are also established in about 200 other major cities. The third Level terminals are directly connected to about 20000 of the most important factories.
What's even more frightening is that there was no fiber optics at that time, and these were all connected by telephone lines.
Although the computing performance of the T55 is now stronger than when the OGAS system was planned, and optical fibers can be mass-produced in half a year, Tang Hua still does not intend to engage in this set of things.Because the size and complexity also means that the price is astonishing, Tang Hua suspects that the price issue is the main reason why OGAS was eventually rejected by the Soviet high-level.
The "National Computing and Information Processing System" has set up a computing center in the Moscow State Planning Commission and four sub-centers in Minsk, Kiev, Novosibirsk, and Vladivostok.
The first phase of the project is here first.
The production orders obtained after statistical calculations by the Moscow center and the regional center can be sent to the factory by running errands or by telegram.
The cores of computing centers and sub-centers are all T55M medium-sized computers.The connection between the center and the sub-center is carried out by modem + telephone line.
The architectural design drawings were copied, produced, packaged, and sent away by secret mail from the Soviet embassy.
Tang Hua called and specifically asked about the production line of the Computer Institute and the factory in Wangjing.
"In the first half of 1955, we may need to supply 40 T55Ms," Tang Hua said. "Be prepared for this."
Supercomputer, T55M, T55L... Dedicated computer...
Tang Hua is now very much looking forward to the start of production at the chip factory in Shanghai. This is to ease the pressure on production. Second, it is time for Group X’s laboratory in Beijing to be freed from repetitive mechanical production to carry out the next stage of research.
……
Supercomputing Center of Electronic Information Industry Bureau.
A total of two TC55s were produced, one in the Ministry of Water Resources and one in the Supercomputing Center.The third one is being manufactured, and it is scheduled to be sent to the supercomputing center of the Institute of Atomic Energy Sciences (Institute of Modern Physics) after leaving the factory.
Today, the supercomputing center encountered a very rare incident. The person in charge of the center called Tang Hua's office for help.
"Minister Tang, I have been here since the establishment of the Supercomputing Center, and this is the first time I have encountered this situation," Lao Liang, the person in charge of the center, called. died."
Tang Hua: "Is there something wrong with his program?"
Old Liang: "There is no problem. We have checked his program. The computer is not faulty, but the amount of calculation is too large."
Tang Hua rode his bicycle from the area of the Institute of Communication and Electronics in the computer compound to the supercomputing center of the Institute of Computer Science.
"Minister Tang, this program has been running for 4 days. This morning we verified what step the calculation has reached, and it turned out that the calculation is 31% complete," Lao Liang said. "And when computing, all the storage space is taken up."
4% in 31 days, which means that it will take 12 to 13 days to complete the calculation.
In the past 4 days, the computing power of the TC55 supercomputer has been fully occupied by this program, and no other programs have been run simultaneously.Other people who apply for computing can only use the TC54 next door. ——Although it is a generation older, it is still very fast.
从去年TC54摆在这儿开始，有资格使用TC54的单位，来申请做的计算都很少超过1天的量。大多数人提交的计算会在10分钟到2小时之内拿到结果——一台TC54的运算速度可是至少相当于100台用集成电路的X54，TC55相对于T55也差不多。
There are even some programs with very small amount of calculations, which can be completed in less than 2 minutes. The person who submits the calculation application goes downstairs to smoke a cigarette, and comes up to see that his calculation needs have produced the results.
"What computing needs, let me see." Tang Hua found a console with a monitor beside the TC55.
"Minister Tang, it was the two of us who submitted the computer application. It's causing trouble for your work." A voice behind said.
Tang Hua looked back. "Hello, Director Tu. It doesn't matter, we still have a TC54, which can be used by other users who apply for computing."
"Minister Tang, I think this calculation process can be stopped."
Behind Tu Changwang, Zhu Kezhen flashed out.
“我要计算的是华北地区10~14天的中长期天气预报。没想到TC55要12~13天才能得出结果，那，这个计算也就失去意义了。”
Chapter 79, imitation AK
Tang Hua saw that the person who came was Director Tu of the National Meteorological Administration, and he understood it in his heart.
In December 1949, the Meteorological Bureau of the Central Military Commission was established. In August 12, it was renamed the National Meteorological Bureau and changed from the Military Commission to the Government Administration Council.Zhu Kezhen is now an expert consultant of the National Meteorological Administration.
1944年6月4日晚上9点半，英国的气象专家找到盟军霸王行动司令部，说诺曼地区天气即将转好，6月6日白天会有12小时的好天气。司令部里的英美海陆空指挥班子看看窗外——风雨交加。
At 5:3 a.m. on the 30th, meteorological experts reaffirmed that the weather in the Normandy area was good from the early morning to the daytime on June 6 after checking calculations.Allied Commander-in-Chief Eisenhower looked out of the window, and it was no longer stormy, but pouring rain...
Eisenhower considered for a moment, ordered the fleet to continue operations, and the Normandy landing battle started as scheduled.
The weather forecast of the Allied Forces before the Battle of Normandy is a legend in the history of warfare, but meteorological experts at that time made short-term weather forecasts for 24 to 72 hours.There are weather stations with sufficient density and coverage, and there are theories. It is enough to calculate by hand + simple calculator, plus a little experience and intuition of meteorological experts.
The medium and long-term weather forecast of 10 days to 15 days is not the way of short-term weather forecast.In addition to weather stations, theory, and intuition, one must also have strong computing power, capable of completing calculations of hundreds of thousands of data that require countless iterations in a short period of time.
When the Yinhe-1 billion-time supercomputer was successfully developed, Xinhua News Agency reported that it had three uses: oil exploration, weather forecasting and engineering physics research.
Meteorological forecasting means that the National Meteorological Administration hopes to establish a medium and long-term weather forecasting system in the 80s.
Engineering physics research refers to nuclear weapons research.
However, what Deng Jiaxian wanted was computer simulation of nuclear explosions. For this purpose, the computing speed and storage capacity of Yinhe-1 are not enough, and Yinhe-1 uses imported chips.
Deng Jia first reported to the central government, requesting to continue to develop stronger supercomputers, and then Yinhe-2 and Yinhe-3 were born.
Shortly after the launch of the Yinhe-3, China announced the cessation of nuclear testing.
……
Let’s go back to the medium and long-term weather forecast system.
The current TC55 is not enough to calculate medium and long-term meteorological data.According to preliminary estimates, the computing power will probably increase by an order of magnitude, and the memory and storage-processing architecture may also need to be changed.
"President Zhu," Tang Hua said to Zhu Kezhen, "I'm sorry, but the computing power of the current computer is still relatively limited. The amount of data calculation for the medium and long-term weather forecast in the region should be too large, and the results cannot be calculated in a short time. "
"No, no, no," Zhu Kezhen waved his hand, "We are the ones who have caused you trouble. This is actually an idea I conceived last year, using a method that relies more on numerical calculations to obtain ten-day weather forecasts."
Not only is it a 10-day ten-day (medium- and long-term) weather forecast, but Zhu Kezhen later proposed monthly weather forecast (long-term) for the first time in China.
Tang Hua: "There is also a supercomputer in the supercomputing center. Although it is a slightly older model, it can also be used to meet the application of large-scale calculations. Your weather forecast process should keep running until the results are obtained. Although there is no timeliness, but it can be compared with the real weather to see its accuracy."
Zhu Kezhen: "Then trouble the Supercomputing Center."
"Mid- and long-term weather forecasts are crucial to both agriculture and the military. We will definitely develop computers that are fast enough to carry out weather forecast calculations across the country," Tang Hua said. "But the development of computer technology takes time. Please wait for three or five years." Year."
Zhu Kezhen nodded, "The speed of your current computer is more than twice that of last year's. Three to five years later...if it still develops at this speed, maybe it will be possible."