Rebirth: Fight For Life

  Chapter 590 Operation of Atmospheric Motion Equations for 40 Billion Grid Columns

   After sending Gu Tao and Bao Jiaqi away, the project team went to Gu Song’s courtyard, had a good meal, and then began to have a holiday.

   After the holiday is over, it is to connect with the meteorological department.

   First of all, the data forecasted every day is compared. The calculation speed, budget capacity, and accuracy will be seen by the veterans.

  Then, when the big cold wave before the Spring Festival is forecasted earlier, it should attract more attention, right?

   In the past six months, others saw the results of the Flint Research Institute and felt that the speed was very fast, but Gu Song was actually quite relaxed.

  08 is a critical year.

   For this reason, in the underground laboratory of Xiangshan Bieyuan, there was almost no stop.

   and Xie Yinran were tired for two days, and she entered the final exam state again.

   And Gu Song concentrated on preparing for the first launch on Flint Island.

   As a business, there is absolutely no need for his launch site to wait until it is built to specifications before it is put into use.

   Besides, this launch has a special mission.

Therefore, after more than half a year of intensive construction on the Flint Island and the rough construction of the launch site, the first custom Candle Dragon rocket provided by the Aerospace Science and Technology Group, as well as the "test satellite" that Gu Song tinkered with in the Xiangshan Courtyard. The key parts reached the island by sea.

   Then, at a high cost, he was invited to the Flint Island to help the team that carried out the rocket hoisting and testing.

   Ye Ziheng also went to the island to show his attention.

   Looking at the rough space launch site, he said helplessly: "Why do you have to come here to launch? Ocean shipping, there are many variables, and now the foundation here is poor."

   Jian Yushu didn't care: "We always have to enter the next stage. The team trains with you every day, and we also need to go through some practical projects to accumulate experience."

   Ye Ziheng shook his head: "You are really willing to spend money."

  The normal course is to learn, boil and accumulate experience year by year. For such a long time, the salary package of these people is very ordinary.

   is another form of cost reduction, in exchange for time.

   On the other hand, Flint Aerospace is the opposite, preferring to exchange money for time, but also to allow the team to grow rapidly.

   He looked at the busy crowd in the hoisting workshop, and the candle dragon rocket, praying that it could successfully complete the task at that time.

   Although he was very worried, the level of the team is really not as good as other launch sites, even if there are problems, it is understandable.

   But after all, this is the first Candle Dragon rocket delivered by the Aerospace Science and Technology Group. If this is the case, it will fail directly. How much pain?

   came out and looked at the seamless steel rail on the ground, Ye Yeheng became more and more emotional: "You guys are doing aerospace, that's really serious!"

   Jian Yushu smiled arrogantly, of course.

   Talking about the goal, now Flint Aerospace is hoisting the world.

   Not long ago, in November, SpaceX successfully launched the test again, blowing loudly, saying that the next step is to deliver cargo to the International Space Station, breaking through the docking technology.

   Jian Yushu, this guy can't wait for the time to speed up, so he hurried to the next stage.

   Helpless, this is the reality.

   Flint Aerospace is too showy to make the whole world nervous.

   He looked at the busy preparations for the launch team, thinking about the mission of this launch mission, and sighed inwardly.

  …

  Yanjing, an expert from the meteorological department, Kou Xun, has been added to the quantum computer project team.

   He's been in shock these days.

   He never imagined that he would see a quantum computer in his lifetime.

   It's not the kind that the media blows up, it's the kind that has real objects and has played a role in a field that he can understand.

   Under the instructions of the leader, the weather data saved over the years was brought over.

  Public channels, the global meteorological data that can be found are also collected as much as possible.

   Then, while the weather forecast prediction algorithm currently used by the meteorological department was compiled into the operation program of the quantum computer, a group of people used artificial intelligence to continuously iterate and prepare the optimization algorithm with the help of Zhang Qingping.

The first wave of    hit Kou Xun, in fact, the most intuitive, that is the speed of operation.

   Weather forecasting is numerical weather forecasting.

  Simply put, after years of research, scientists have summed up six equations of atmospheric motion. The temporal and spatial variations of the wind, pressure, density and temperature fields in the atmosphere can be described by this set of equations.

   In fact, these equations are all under idealized conditions.

   Just like vacuum and non-vacuum, smooth and frictional, to really calculate and analyze these equations, parameterization is required.

   In addition to meteorological observation data as initial conditions for analysis, there is also an analytical scale.

   This scale is a bit like resolution.

  The atmospheric environment is figured out, and all changes are related. To calculate, then draw the grid first.

   First, divide the surface from the surface to the top of the atmosphere into N layers.

   Then, on the surface, also divide, a grid length and width X kilometers.

   In this way, you get, um, the surface area of ​​510 million square kilometers divided by the square of X, which is about millions of grids. Multiply millions by N, OK, this is the grid point that needs to be calculated for tens of millions of millions.

   Then, it depends on whether you use 10 minutes, 30 minutes, or a longer time to calculate the value change in each grid.

   If it is 10 minutes, make a 10-day forecast. Well, then each square counts 1440 times.

   The real example is calculated with a resolution of 15 kilometers in length and width of each grid. It is calculated twice a day, and it takes two hours for a giant supercomputer to complete the calculation of the equations of atmospheric motion in a column of 40 billion grid points. Then, the numerical changes provided are used as a basis for professional weather forecasting.

  Partial differential equation + floating point calculation + n grid points in the horizontal direction + n layers in the vertical direction + time step limit + parameterization scheme +... super calculation is very tiring...

  The calculation also takes into account the issue of timeliness. The result was just calculated today, but the sun has gone down and the daylilies are cold. Who cares?

   Weather forecasting is such a tangled process. Predictability is rooted in the interaction of atmosphere, land surface, vegetation, sea ice and sea surface. And there are always unpredictable small-scale factors that are related to the "chaos" theory.

   At present, it is very painful to achieve a resolution of 12 kilometers at the limit, and it is even ridiculed that the speed of calculation cannot keep up with the speed of the actual change of the atmosphere.

   For example, at a resolution of 1 km, the changes caused by special factors such as small cloud clusters cannot be included in the calculation amount at all.

  The reason why Kou Xun was so shocked was the time it took for a quantum computer to complete one calculation of the same amount.

  17 seconds.

  In the data brought over, the results of many supercomputing calculations in the past are saved for comparison.

   Kou Xun was dumbfounded there, while Zhang Qingping, Guo Yue and the others were discussing in whispers.

   "Where is the problem, why does it take so long?"

   "Should the error correction mechanism be optimized specifically for this model?"

   "Old Zhang, look at it, are there any problems that can be transformed in a way that is more suitable for quantum computing?"

   "I think another way of thinking is to increase the resolution and use more data to map each other, which will improve the fidelity?"

   "In short, there must be a problem with this time. 100 qubits! If the supercomputing time is about 2 hours, how can it be within 10 seconds here?"

   "Take a walk, analyze the data, and quickly change it before recalculating."

   Kou Xun, who was holding the result, looked dazed.

  17 seconds, why are you so upset?

  Who am I, where am I, what happened?

   (end of this chapter)