The Red Era, the Opening University Speech

Chapter 252 Three Future Industry Leaders

"So what is this routing selection and router? Is it the path for data transmission?" the young team member continued to ask.

Others also looked inquiringly.

Faced with these new terms, many of them have only a vague understanding of them.

Li Mu said patiently: "Routing selection is a mechanism that ensures that packet information is transmitted to the destination in an efficient and timely manner."

“It is the process by which routers forward information from one place to another based on certain path selection and data forwarding strategies.”

"What about Transmission Control Protocol? What is this?" Another young team member raised his hand.

Someone followed up and said, “And how should we build this protocol…”

……

Time passed quickly as we asked and answered questions.

Li Mu spent a whole morning before everyone could barely "understand" what he said.

Of course, that’s what they claim.

It is obvious that there is not much credibility when someone says this while holding a notebook that is almost half a book thick.

Therefore, after finishing his meal, he stayed in the network protocol stack R&D group for another hour before going to other Type 01 operating system R&D groups to check on the situation.

They also encountered many problems.

On the kernel development team side, Professor Zhang Aihua said as soon as he saw him coming: "Advisor Li, it seems that our original computer hardware architecture cannot be well adapted to the kernel you provided."

"It would be very difficult to coordinate and manage the hard disk access, network card transmission, and external device input you mentioned."

"Don't worry." Li Mu waved his hand, then went straight up and said, "To research what kind of kernel, we must first clarify our needs, which is to make a high-performance computer."

"Based on this, we can imagine a design architecture that concentrates the main functions of the operating system in the kernel. I call it a macro kernel."

"As a result, it has high execution efficiency, good resource management capabilities, modular design and system API functions."

"Of course, with so many functions, the amount of code required will increase significantly, and if a module has a flaw, it will affect the stability of the entire operating system."

"If you think it's too difficult, you can choose another microkernel solution that only implements the most basic functions, and other functions are implemented through inter-process communication."

After he finished speaking, he looked at Zhang Aihua.

The latter glanced at the team members behind him and said immediately: "Advisor Li, is there any need to say this? We must research a better macro kernel."

"That's right, if we're going to do it, we'd better do it!" The others also shouted.

Hearing this, Li Mu was a little amused and said quickly: "Don't get me wrong, macro kernels and micro kernels each have their own advantages and disadvantages, there is no good or bad distinction."

"It's just that the macro kernel is more suitable for our current computer development."

"We will still study microkernels in the future, so it doesn't matter which comes first."

After he finished speaking, everyone nodded, but their eyes were full of fighting spirit, and it was obvious that they still intended to study the macro kernel.

Li Mu did not continue to persuade him upon seeing this. It was a good thing that his men had confidence, and there was no need for him to pour cold water on them.

Immediately afterwards, he went to the interactive command interpreter development group and the file system development group to help them solve the problems they encountered.

……

It was almost time for lunch when Li Mu came out of the file system R&D group, accompanied by several young researchers.

The file system is a relatively simple part of the Type 01 operating system architecture, so when he divided the team into groups, he chose young people as the team members, along with Lin Guosheng, an old professor from the China Europe International Business School.

The compilation of textbooks for the software engineering major is still in full swing, and one day in the future it will surely be implemented in Kyoto and even nationwide.

By then, we will need enough talents to meet the educational and teaching needs of various places.

Of course, the people in the Semiconductor Research Institute are all Li Mu’s treasures, and he would not let them go easily. At most, he would send them out to run training classes and give lectures.

Just as Li Mu finished answering all the questions and waved goodbye to several young researchers.

At the other end of the corridor, Wu Youxiang suddenly walked over with a smile.

He first brought Li Mu to his office, then dropped a bombshell, saying, "The higher-ups are interested in developing a new generation of supercomputers, and I came to ask for your opinion."

"If you think it's feasible, we'll start preparing right away."

"A new generation of supercomputer?" After hearing this, Li Mu fell into a brief contemplation.

As for developing a new generation of supercomputer, it is not impossible. After all, he has already put the development of Type 03 computer on the agenda.

But the problem is not without it.

The previous supercomputer cost more than 8 million, and this cost made Li Mu feel heartbroken just thinking about it.

Although technological advances will lead to a reduction in the cost of integrated circuits, the overall cost of supercomputers will never go down.

High-end processors, memory, storage and other components all require a lot of research and development costs.

"Of course you can." Li Mu hesitated for a while and agreed reluctantly.

Wu Youhao nodded slightly, without rushing to speak, waiting for the "but" that followed.

Li Mu continued, "But it's too expensive. I don't think it's worth it. My suggestion is to wait a little longer until the integration level exceeds 10."

"It doesn't matter if it's expensive. Money is only money when it's spent where it should be spent. As long as it can be done, the country will definitely give its full support!" After hearing this, Wu Youxiang was a little amused.

He thought it was some kind of difficult problem, but it turned out that Li Mu had been struggling for a long time because it was too expensive.

No matter how high the cost is, for the sake of the country's scientific and technological development, they will grit their teeth and tighten their belts to save money for the new supercomputer.

Li Mu said helplessly: "I know the country will definitely support it, but I just don't think it's necessary. The current supercomputers should be able to meet our needs."

As soon as he finished speaking, he noticed that Wu Youxiang had a strange expression on his face.

The latter said with a smile: "You really don't know the situation. If you want to use the supercomputer, you just need to tell me and Lao Wang, but there are thousands of research institutes across the country, and each of them has to wait in line. Now it's almost June or July."

"Are you so nervous?" Li Mu was startled.

He only knew that the supercomputer ran almost 24 hours a day, and he never noticed how many people were waiting to use it.

Wu Youxiang said:

“That’s how nervous I am!”

"Okay, since you think it's worth researching, I'll just tell the higher-ups directly."

"I think a team of experts will be here soon, so you should prepare well."

"Okay." Li Mu nodded and agreed.

If it is for research purposes, then the cost issue will not be taken into consideration.

...Afterwards, Li Mu and Wu Youxiang said goodbye, went to the canteen to have dinner, and then rode their bicycles back home.

When he thought about the computing tasks scheduled until June and July, he walked much faster than usual.

After arriving home, he locked himself in the study and began designing a new supercomputer.

The upgrade of supercomputers mainly involves the optimization of processors, memory, storage systems, network connections, and software and algorithms.

As the core, the processor can significantly enhance the computing power of supercomputers, and the improvement method is simple and crude, such as a more powerful CPU or more advanced architecture technology.

There is no way to improve the integration level, Li Mu can only look for breakthroughs in architectural technology.

If we want to improve the memory and storage systems, we can only research more advanced random access memory and high-performance solid-state drives, but these two technologies are not easy to develop.

Software and algorithm optimization is difficult to achieve significant improvements given the current level of software engineering.

Li Mu set his sights on the final network connection upgrade.

With the Internet, supercomputers can achieve cross-platform collaboration, and the computing speed will definitely be greatly improved.

Of course, how to achieve high-speed network connections to support large-scale data transmission and processing is a big problem.

The network speed of a supercomputer is mainly determined by the network structure connecting the CPU and GPU, the Internet, the storage system, the system software, and the application software.

The network connecting the CPU and GPU in later generations mainly relies on PCle bus and NVLink technology.

PCIE, the full name of PCI-Express, is a high-speed serial computer expansion bus standard proposed by Intel in 2001.

However, this technology is difficult to implement, so Li Mu plans to develop its predecessor - PCI.

It is used to define the local bus standard, which allows up to 10 expansion cards that comply with the PCI standard to be installed in the computer.

In terms of transmission speed, PCI can reach up to 133MB/s. Although it is far less than PCIE's 4GB/s or even 8GB/s, it can already meet the needs of the next generation of supercomputer development.

NVLink is a technology developed by NVIDIA for point-to-point high-speed interconnection between GPUs. It aims to break through the limitations of PCIE interconnection bandwidth and achieve low-latency, high-bandwidth data interconnection between GPU chips, allowing GPUs to work together more efficiently.

However, this technology appeared quite late, and it is unknown whether it can be developed.

Li Mu plans to use it as PLAN-B, study it first, and if it works, continue. If it doesn't work, he can accumulate experience and restart it at the right time.

Apart from the Internet, the remaining three parts, storage system, system software and application software, can be explored by Huang Kun and Wu Xi by following their own example.

Of course, this is still quite difficult for them, and he needs to provide guidance in the right direction through the plan.

Although the Internet can directly improve the computing power of supercomputers by increasing their number, since Li Mu has decided to develop it, he must strive for perfection.

The storage of future supercomputers will mainly use physical storage devices such as memory sticks, hard drives and flash memory. Virtual memory technology will also be used to allow programs, data and stacks to exceed the actual physical memory.

Li Mu didn't dare to think about virtual storage technology, but he could try memory sticks, hard drives and flash memory.

There are also system software and application software...

……

As he wrote, Li Mu suddenly couldn't help rubbing his sore neck.

When he looked up, he found that it was already 2 o'clock. There was a cup of cold boiled water on his left hand. Wang Ya probably saw that he was too engrossed, so she put it down silently and turned away without disturbing him.

He rubbed his brow, put away the written documents in front of him, and then planned to go back to sleep.

At this moment, a message alert sounded in his mind.

"The frequency of these messages is quite frequent!" Li Mu smiled in his heart and opened the chat page.

Ask: "Li, what should I do? She invites me out tomorrow. Should I go?"

"Of course you should go. Spend her money lavishly and she will definitely hate you." Li Mu replied quickly.

He felt like he was back in college, having to act as a military advisor to his roommates from time to time.

Ask: "Spending her money? That's not good. Don't say we're not a real couple. Even if we are, this is not right."

"That's why she will leave you immediately. Believe me, it's okay to make a mistake. At worst you can return it to her later." Li Mu said with a smile.

When Ask asked this question, his goal was achieved.

It was originally a question of whether to go or not, but now it has become a question of whether to spend the girls' money if I go.

The first move he made was a pure gold Blue-Eyes White Dragon worth tens of thousands of dollars. Li Mu estimated that the other party's family background should not be poor and he would not care about the little money Ask spent.

Besides, based on his understanding of Ask, he would not take advantage of him for nothing and would definitely buy something else in return.

After a few tries, the thing will be done.

……

The next day, early.

After washing up and exercising, Li Mu walked out of the special building yawning.

Arriving at the Semiconductor Research Institute, he was about to jog up the stairs when someone suddenly called him from behind: "Li Mu."

Hearing the sound, he turned around.

Wang Shoujue walked forward with a smile, followed by three unfamiliar faces. They seemed to be newcomers to the institute. They were not very old and their eyes were full of nervousness when they looked at him.

"Good morning, Director Wang. This is..." Li Mu greeted them and then looked past Wang Shoujue and at the three young men.

Without waiting for Wang Shoujue to introduce, the three of them said loudly:

"Hello, Consultant Li, my name is Yang Tianxing."

"Hello, Consultant Li, my name is Zhou Xiling."

"Hello, Consultant Li, my name is Su Dongzhuang."

"Hello, hello." Li Mu was startled by the three people, especially the fiery look in their eyes, which made him a little confused.

Soon, the doubt was resolved.

Yang Tianxing said excitedly: "Advisor Li, we have long heard that you are the inventor king of our Xia country. I didn't expect that you are also the chief designer of our country's first generation of supercomputers!"

"Yes, no wonder you are called the 'leader of young scientists in our country'. It turns out that you have such an achievement that no one knows about!" Zhou Xiling also said.

Su Dongzhuang's eyes were filled with excitement, and he said, "Advisor Li, please give me more guidance in the future!"

……

The three excited people did not hide their excitement at all, and while they were talking, they continued to introduce themselves in depth.

Yang Tianxing graduated from the Computer Department of China Europe International Business University and previously worked at the North China Institute of Computing Technology.

(End of this chapter)