Chapter 362: The most powerful brain, modular parallel program development!
After chatting some more about the project, Lin Zhi took out a beautifully decorated small box from his bag and placed it on Chang Haonan's table.
“By the way, Teacher Chang, we also brought you a gift.”
A model racing car.
Indianapolis has a long history of the Indy Formula Series and a rich racing culture.
Originally, she was going to bring up the topic when a few people went to see the Indianapolis Motor Speedway, but Chang Haonan interrupted the progress of the chat.
Having no choice but to give it a bit stiffly.
“IndyCar?”
The gift packaging box is transparent, and Chang Haonan can see the contents inside without opening it.
“Do you even know about this game?”
Song Tingyu's eyes lit up:
“Before I went to the United States this time, I had only heard of F1.”
“Well...I also like F1 better.”
Chang Haonan looked at the exquisite racing model for a while:
“And I guess Michael Schumacher will win with Ferrari, more than one.”
After saying this, he noticed that the corners of Song Tingyu's lips curled up involuntarily, obviously disapproving of it, but did not refute out of politeness.
After all, the Red Circus has not won the driver's championship in almost 20 years. F1 fans in this era are more concentrated on Williams and McLaren.
Of course Chang Haonan just said it casually and did not continue to dwell on this issue.
He placed the model next to the computer monitor:
“You should make preparations after you go back, and you can also listen to Academician Yang's opinion. In short, if you join my project team, you will definitely not be able to go home during the summer vacation.”
As a newly confirmed national key laboratory, the unveiling ceremony was naturally quite grand, even bigger than the previous time when the supercomputing center was established.
This is not even a request from Chang Haonan, but an arrangement made by the school.
In any case, this laboratory is built on the campus of Beijing Airlines after all, and that's a shameful thing.
In addition, Xu Yang and Yao Mengna also rushed back from Haojing.
The former will be responsible for part of the software development work, while the latter will specifically optimize the aircraft design based on actual use experience.
Since Chang Haonan has accepted the title of chief scientist, the laboratory does not have a director for the time being, and only two deputy directors are responsible for assisting the work.
According to regulations, the State Key Laboratory will conduct an acceptance inspection 2-4 years after its completion.
As long as it is signed by the director at that time, it will be fine.
Considering Chang Haonan's growth rate, he should have graduated by then, and there will definitely be no problem in officially assuming the position of laboratory director.
So there's no rush.
Considering the tight time, Chang Haonan chose to organize the first meeting of the project team on the night of the unveiling ceremony.
Originally, Chang Haonan was planning to call a work meeting directly and start assigning tasks to start work as soon as possible.
But just as the meeting was about to begin, Song Jian, who was just here for a formality, suddenly arrived.
Following him were Yang Fuqing, Li Jie and several other academicians who were also invited to attend the ceremony.
This is definitely impossible to exclude.
Director of the Science and Technology Committee is just a less important one among Song Jian's many titles.
In fact, he is an expert in optimal control system theory and distributed parameter control theory, an academician of the Chinese Academy of Sciences and the Academy of Engineering, a foreign academician of the Royal Swedish Academy of Engineering Sciences, the U.S. National Academy of Engineering, the Russian Academy of Sciences, the proposer and person in charge of the 973 Project and the Spark Project...
And if nothing else happens, he should also be the next dean of the Academy of Engineering.
In order to show respect for the big guys, the title was temporarily changed to a seminar.
Of course, since it is a discussion, the content of the meeting that has been prepared cannot be changed. He plays the leading role alone, and its essence is closer to answering questions.
Originally, Chang Haonan meant to let Song Jian say a few words first.
But the latter waved his hand to show that he didn't need it, and spread out a notebook on the table in front of him:
"We came here without warning for inspection and learning. Comrade Chang Haonan, just carry out the meeting process as usual. If you have any questions, we will ask them in the speech session later." Chang Haonan quickly understood that the other party was doing this. I'm afraid I came here with the intention of giving myself a platform.
His main contacts in the past year were, after all, units under the Commission of Science, Technology and Industry for National Defense. However, the laboratory established this time involved cooperation with the Academy of Sciences and many universities belonging to different systems, and it may not be as smooth as before.
With Song Jian's words, the situation is completely different.
So after the brief introduction, Chang Haonan didn't waste any time and got straight to the point:
“First of all, welcome to join the Numerical Simulation and Data Analysis Laboratory.”
"I believe that anyone who has done research using simulation methods knows that the current mainstream numerical simulation software in foreign countries is difficult to balance calculation efficiency and accuracy, and it is difficult to obtain results for some complex problems, especially complex problems with multi-field coupling. ”
“So our first project, and also the most important project, is to first create a handy tool for ourselves, which is a brand-new, multi-physics oriented simulation software.”
Chang Haonan turned to the next page of the PPT, which showed the workflow he had spent several days designing.
For this purpose, he spent all the scientific research points provided by the system.
Even so, it can still only provide a basic framework, which will require continuous upgrades and maintenance.
It is enough to show the complexity of this project.
“Compared with the traditional methods in the past, we have mainly improved in the following four aspects.”
"First, it is a new dimensionality reduction method for nonlinear partial differential equations. I have sent the relevant theoretical content to you in writing, and I will be personally responsible for this part, so there will be no details at the meeting. Let's discuss it again. If there is anything you are interested in or don't understand, you can talk to me later.”
“Second, it is to develop new time-frequency domain algorithms for two sets of strong coupling fields: force-thermal and electromagnetic.”
“Third, for most problems that can be solved using traditional methods, we will not completely abandon the finite element method, but for complex problems, a new parallel finite element analysis process needs to be developed.”
“Fourth, it is to optimize the key step in almost all numerical analysis methods, that is, the mesh generation method, especially for some complex 'dirty' geometric problems, we must come up with corresponding solutions.”
“In order to improve efficiency, I will divide the next three tasks into 9 modules and hand them over to different groups for synchronous and parallel development. Finally, they will be centralized and connected to me to form a complete application.”
As soon as this sentence came out, the entire conference room was unsurprisingly filled with exclamations.
Even the mature and steady Song Jian and others raised their heads and looked at Chang Haonan on the stage in surprise.
In fact, the four aspects mentioned by the latter may not sound like much, but in fact they are equivalent to overthrowing and rebuilding the traditional numerical analysis method. The difficulty is comparable to moving a mountain.
Now he is not prepared to do it one by one in order, but wants to do it at the same time?
Do you need to have all the modules connected to him?
This not only requires strong overall control, but more importantly, the fault tolerance rate is very low.
“Mr. Chang, the modular development you mentioned is indeed very efficient, but if any part develops some structural problems, it will affect the whole body..."
“Yes, Mr. Chang, this is good for us. Isn't the pressure on you too great? After all, you have to integrate all the modules..."
Everyone's reaction was not beyond Chang Haonan's expectations.
In fact, without the help of the system, he would not dare to do such tricks.
Lockheed Martin used this method in its previous life when developing the flight control software for the F35.
And it was also outsourced to Indian programmers.
Finally, until Chang Haonan was reborn, he had been in service for ten years, and there were still more than a hundred bugs that had not been fixed...
He raised his hands and pressed them down in mid-air to signal everyone to be quiet.
“I can confirm that there will be no structural design problems. As long as there are no internal errors in all the modules, they will be able to run stably after being assembled together.”
Chang Haonan said in an extremely determined tone:
"As for the bugs that occur during the programming process, this is inevitable, but it will not have a big impact on our project, because after you submit the project to me, if there are bugs, I can give you feedback to the specific location. , just change it when the time comes.”
"In addition, although each of you is only responsible for a part of the program, you must also have an understanding of the overall development ideas, so as to ensure that there are no problems in the direction when performing specific tasks. Therefore, there will be a total of three meetings of the same nature, one each. Introducing the specific content involved in the next three parts of the work, including theory and algorithms. For example, today I will talk about the most basic fourth part, which is the automatic grid generation algorithm.”
“In general, as long as you follow my guidance and express the ideas I have conceived in the form of code, this project will definitely reach the end.”
At the end of a few words, the entire audience fell silent.
If someone else had stood on it, he would have been bombarded long ago.
but now.
On the one hand, it is the support that Song Jian has just shown, and on the other hand, everyone is more or less aware of some of the chief scientist's previous impressive deeds.
Coupled with Chang Haonan's confident but not blind feeling when he just spoke...
An idea popped into everyone's mind almost at the same time:
“Is it really possible?”
Chang Haonan looked at the participants with different expressions in front of him:
“Did I understand it clearly?”
(End of this chapter)
Chapter end
Report
|
Donate
Oh o, this user has not set a donation button.
|