Great Country Academician

Looking at the manuscript paper in his hand, Mao Shun couldn't help but swallow the air dryly.

Although he couldn't understand whether the above answer was correct, he didn't think that the person in front of him would make up some random answers to fool him.

After all, there was no reason for him to do this.

It's just that in just over an hour, they can solve a problem that they have been studying for several years and that many mathematicians have not been able to solve. Isn't this a bit too scary?

Is this the strength of the number one person in mathematics today?

Taking a deep breath, he carefully arranged the manuscript papers in his hands, then took out a document bag from his backpack and put it in carefully.

No matter what, he would take these manuscripts back for study.

After all, chip R\u0026D and design are related to these things. HiSilicon and Huaxin need to adjust their mathematical models and control instructions based on the above answers to design and manufacture chips with better performance.

As for whether the optimization effect can be achieved as expected, we need to wait for the optimized chip to come out before looking at the results.

Opposite him, Xu Chuan didn't pay attention to the changing expressions on Mao Shun's face. He then said, "By the way, there is one more thing."

Hearing this, Mao Shun quickly came to his senses and said quickly: "Academician Xu, please tell me."

Before he came, he was told by his superiors that if Academician Xu made any request during this trip, he would have to find a way to fulfill it no matter what.

Xu Chuan: "When you go back, talk to your leaders and see if you can send me detailed information on the NPU neural network architecture you are studying."

"Mainly because many of your problems are based on the architecture you studied. I need to study and optimize based on your foundation."

Mao Shun nodded and said, "Okay, I'll report it right away."

Xu Chuan: "Okay, then leave this hard drive here for now, and I will study it later."

Mao Shun stood up, bowed slightly and said thanks: "I'm really sorry to trouble you, Academician Xu."

Xu Chuan smiled, waved his hand and said, "It's a small thing."

After sending Mao Shun away, Xu Chuan returned to the study, clicked on these mathematical problems, and started reading them again.

On the other side, Mao Shun, who had just left the villa, left the community under the leadership of Zheng Hai, and then found a quiet place to call to report the situation.

"Well, I understand. You can stay there first. Time doesn't matter. I'll have someone sort out the information and send it to you. You can then talk to Academician Xu." A steady male voice spoke on the other end of the phone. Passed over.

Mao Shun nodded, thought about something, and said quickly: "There is another thing. Regarding the mathematical problems I sent over today, Academician Xu has already solved part of them. The answer sheet is in my hand. You see, Should I send it back, or arrange for someone to come pick it up?"

Hearing this, the other end of the phone was silent for a moment, and then asked with a somewhat puzzled voice: "What do you mean by partially solved? If I remember correctly, you should have just arrived in Jinling today, right?"

Mao Shun: "Literally, I just arrived here this afternoon. I just met Academician Xu. After Academician Xu took the questions from me, he solved three mathematical problems directly on the spot. It took a long time." About an hour.”

After hearing this, there was complete silence on the other end of the phone.

Just when Mao Shun thought that the other party had hung up, the disbelieving voice from the other side of the phone came over.

"you sure?"

Mao Shun nodded and said with certainty: "I have the manuscript paper to solve the problem in my hand. Although I can't understand whether the above answer is correct, I personally don't think Academician Xu will make up some answers to deceive. We can only say that his strength in mathematics far exceeds our imagination."

The person on the other side was silent for a while, and then he said, "Keep the manuscript carefully. I'll ask someone from the Jinling Branch to pick it up from you."

Helping HiSilicon and Huaxin solve mathematical problems in chip design is not something that can be done in a short time, especially if you are not familiar with the overall architecture. Even if you are a Fields Medal winner, it is not that easy.

So after re-reading the answer sheet, Xu Chuan put it aside for the time being and began to collect some information related to chip design and NPU neural network architecture.

In order to solve the problem in a targeted manner, it is necessary to understand the nature of the problem and relevant information.

This is the same whether in mathematics or in other fields.

The days passed like this, and the people arranged by HiSilicon and Huaxin quickly sent over the detailed information on the NPU neural network architecture, and also took away the answer that Xu Chuan spent an hour writing. , seize the time to do research.

Magic City, Zhangjiang Road, Pudong New Area.

In Huaxin International Technology R\u0026D Center, many scientific researchers and staff are busy in this building that does not look very high.

For people working here during this period, leaving early and coming back late, and working overtime are the most common words.

Many people can't even use such words, because they eat, drink, and sleep in this building.

The research and development and design of chips is a very complex task.

Especially the early simulation design work is so complicated that even people in the industry may not be able to understand it as long as they boast about their own field.

If digital IC is like science, then analog IC is more like magic.

Using computers to assist in the design of analog chips is essentially solving high-order differential equations one after another.

From the early design requirements, specification structure, schematic design input, pre-simulation result comparison to mid-term layout design, DRC/LVS rule checking, to later parasitic parameter extraction, back-end simulation, GDS file formation, to the final software Verify whether the desired functions and performance can be achieved, etc.

Every step involves countless processes and the hard work of countless people.

In the laboratory, Yu Mingzhi, the vice president of HiSilicon and the general manager of HiSilicon's chip design department, was sitting anxiously in his office and waiting.

This year is the first time the optimization results will be released after the end of the year. It is also the first time they will solve design problems based on the problem and get results after getting part of the answers from Academician Xu.

Whether it can greatly improve the parameter performance of the chip as imagined; or whether it will be the same as before, with basically no big breakthrough, depends on today's results.

So even though he had taken the position of vice president of HiSilicon, Yu Mingzhi was still a little restless.

After all, this time's chip design and manufacturing is not only about whether China can have its own seven-nanometer, but also about the next series of deployments by HiSilicon and Huawei.

Such as mobile phones, computers, smart fields, etc.

Not to mention other things, in the field of mobile phones alone, they have lacked a chip of their own for too long.

After Qualcomm banned the export of top chips to them a few years ago, Warwick's top flagships once fell into a situation where no chips were available.

Last year, both Qualcomm and TSMC stopped supplying chips, which directly put them in dire straits.

Although Warwick and HiSilicon can live well without the mobile phone business, the feeling of being stuck in the neck and allowed to be manipulated by others is really frustrating.

In the office, Yu Mingzhi flipped through the mailbox and saw some other emails, but he had no intention of solving them, and his eyes drifted to the door from time to time.

"I don't know if Academician Xu can solve this trouble, but apart from him, there is no other more suitable candidate in the country."

"If it still doesn't work, it may be difficult to achieve a breakthrough of seven nanometers in a short time."

Thinking about it, Yu Mingzhi couldn't help but take a deep breath.

Along the way, he knew very well how many difficulties and obstacles HiSilicon and Huaxin had experienced. Even though they had now seen a glimmer of hope, there was still hope that the bottomless abyss would stand in front of the finish line.

If there is a reward in the end, then it will all be worth it.

If nothing comes of it in the end, he won't even know what the future direction will be.

Not just for Warwick, but for the entire country of China.

After all, HiSilicon and Huaxin are almost one of the top combinations in the field of chip R\u0026D and design in China.

Perhaps Ziguang Changdian and the others are studying different routes, but it is really too difficult to make a breakthrough in this area.

When I was lost in thought, there was a knock on the office door.

Before he could even open his mouth knowingly, the door was pushed open with a bang, and another middle-aged man rushed in quickly with a report sheet in hand.

"Mr. Yu, the simulation test results are out!"

The middle-aged man who hurried in had excitement and excitement on his face, and he didn't care about anything else and directly opened his mouth to report.

Yu Mingzhi knew there was nothing else. After hearing the report, he stood up suddenly and asked quickly: "What's the result? Let me see!"

"The overall performance index of the single module task has increased by 24.71% on the original basis! The signal-to-noise ratio, distortion, and filtering capabilities have all been slightly improved!" Pei Zheng, who broke in, reported excitedly and handed over the report results in his hand. Passed it over.

Yu Mingzhi quickly took the results and started reading them.

He is the vice president, but he is more technical, so he can naturally understand the test reports.

Looking at the test values ​​​​of the above items and the original comparison data, the tense face gradually relaxed, and a smile appeared at the corner of the mouth.

"beautiful!"

Clenching his fists tightly, Yu Mingzhi let out a low growl. All the depression and worry he had been feeling for a while were released at this moment.

Although it is only a phased result, for them who have long been stagnant, this is the dawn of hope!

Opposite me, Pei Zheng, who was in charge of the simulation test, also had a bright smile on his face and said, "Yes, it's really beautiful!"

After a slight pause, he continued: "The simulation design data this time, according to the design department, should be able to be optimized and improved in the future."

"They said they are still studying some things. If they can understand it thoroughly, based on the current results, the overall performance index of a single module task will increase by more than 30% on the original basis."

Yu Mingzhi said with a smile: "Well, it's a very good result. I originally expected it to be about ten to fifteen percent, but I didn't expect it to more than double."

Pei Zheng asked curiously: "Such a big breakthrough should be due to progress in the previous key mathematical design problems. Which big name have you poached from abroad? You can solve some of your problems so quickly."

Yu Mingzhi smiled and said, "I'm afraid you have to ask the big boss yourself about this question."

Although he knew it, it was impossible for him to reveal it.

Pei Zheng did not continue to ask questions. All the people who could work in this building had signed confidentiality agreements and received confidentiality training.

Although they are a little curious, they don't need to ask the source of some things, they just need to do their job honestly.

Nodding, Pei Zheng said: "Okay, I've reported it to you, I have to go back to work first."

Yu Mingzhi smiled and said: "Well, I look forward to your next good news."

Pei Zheng left the office, and Yu Mingzhi picked up the simulation design test report on the table. Looking at the excellent data on it, he couldn't help but sigh.

"As expected of the famous Academician Xu, he is amazing."

One hour can solve the problem that has troubled them for more than a year, and the changes brought to the simulation design directly increase the overall performance index of the single module task by 24.71% on the original basis.

This is before their design department has fully understood it. If they fully understand it, it can even increase by about one-third.

This number doesn’t sound like much, but it’s actually terrifying.

After all, this is just a simulation design, and only part of it is solved.

If all the problems they listed in the past can be solved, how much better can the performance of the chip designed this time be improved?

To be honest, he couldn't imagine it.

If that person can participate in chip design, let alone seven nanometers, maybe five nanometers and three nanometers may be able to try to make breakthroughs.

PS: There is another chapter tonight, please vote for me