Chapter 465: Choice of Process
Speaking of which, over the years, Xu You's fastest way to improve his brain power is to rely on improving his systematic learning level.
This level of improvement can instantly gain a mysterious energy, directly raising Xu You's brain power to a higher level.
Of course, this incredible method, which even Xu You himself cannot explain now , should obviously not be suggested to Lou Song.
"My personal experience is that every time I think deeply, I can improve my brain power by a small amount. The brain seems to be like a muscle. The more you exercise it, the stronger it becomes."
Lou Song nodded vigorously and said, "Yeah! I feel the same way. Every time I think about a difficult problem, I feel that my brain changes a little. However, it has to be a very difficult problem. When I was in school, the difficult problems that my teacher left for me had basically no effect."
Speaking of this matter, Xu You couldn't help but laugh.
At that time, Xu You’s roommate Guo Haoran also received special treatment from the teacher because of his “outstanding performance”.
However, those topics that were not worth mentioning in Lou Song’s eyes were full of “good things” in Guo Haochen’s eyes.
"If we want you to find the questions difficult, then they are not ordinary questions. Lou Song, besides normal thinking, do you have any other secrets to improve your brain power?"
Lou Song thought for a moment and said, "It's also solving problems, but the method may be a little different. Sometimes, I can solve the problem directly in my mind, and the speed is faster than using draft paper. Solving difficult problems in this way will be more comfortable than the normal way."
When Xu You heard this, wasn't this similar to his own deep thinking state and brain simulation?
"Lou Song, I feel the same way as you do. When I'm in a good state, I can simulate various reactions in my brain, or play chess with myself. When the brain is thinking at a high level, the speed of improving brain power will indeed be faster."
"Great! I've always felt that the brain is like physical fitness. It's not just innate, it can be improved through training. But there's an age limit. After a certain age, the brain will reach its peak and can no longer be improved. Before reaching this limit, we must seize the opportunity to improve our brain power. I estimate that after the age of 40, it will be difficult to continue to improve."
In the sports world, athletes often peak in their twenties or thirties, and it is rare for them to be over forty and still be able to improve.
The Fields Medal, the "Nobel Prize" in mathematics, is only awarded to mathematicians under the age of 40, which can also illustrate a similar problem.
After Xu You and Lou Song exchanged their methods of improving brain power, they continued their research.
…
With the joint efforts of Xu You, Lou Song and others, as well as Suanjing Artificial Intelligence.
We have successfully completed the development of computer chip EDA software and are ready to enter the chip design stage.
"I never expected that AI could play such a big role in the EDA field. Without artificial intelligence, we would definitely not be able to complete the development of computer chip EDA so quickly and so well."
"Yes, if this continues, there will probably be no more work for us humans in the future, and all work can be solved by AI."
Hearing everyone talking about artificial intelligence, Xu You went over and explained with a smile:
"In fact, AI can help us solve these problems now, largely due to its powerful computing power. In the field of EDA, AI has two main applications: one is predicting performance, and the other is helping tools to be used more automatically. Both of these are inseparable from AI's special algorithms and powerful computing power."
"Professor Xu, can you explain it to us in more detail?"
Seeing that everyone was curious about the specific principles, Xu You began to explain patiently.
“In terms of predicting performance, the backend design process is very long, so what we can see right now is actually very far away from the backend. Relying on the powerful computing power of AI, we can quickly see the performance of the backend, greatly improving the accuracy of the simulation and the performance of the optimization. Even if it can only be improved by 5% or 10%, it is a huge improvement for the overall situation.”
"As for the issue of tool automation, it is easier to understand. If you only run the data once, there will be no difference in speed between manual and AI. If you run it dozens or hundreds of times in a row, the speed difference will be huge. In this regard, our AI is not only faster, but also gives recommended parameters and expected results within a reasonable range."
Xu You used very simple language to explain the specific principles to everyone, and everyone quickly understood it.
The next step is to use EDA software to design the chip.
Because of the compatibility issues between chips and systems, a mature system can only be made after the chip design is completed.
It is very normal for various bugs to occur between chips and systems.
The system can be modified to accommodate the chip, but once the chip is made, it cannot be modified anymore.
There is also active discussion on
"According to the latest internal information, our EUV lithography machine has basically completed all preparation work, and only the final testing stage remains. If nothing unexpected happens, we will be able to make our own high-precision chips soon."
This news makes everyone very excited, but there is no conclusion yet on when this lithography machine can be used .
"My suggestion is to design computer chips with 10nm or 14nm process technology first. Such process technology does not require the use of EUV lithography machines, and our technology is more mature. Moreover, compared with mobile phone chips, computer chips do not have such high requirements for process technology. As we are just starting out in this industry, we still need to start with 10nm or 14nm process technology, and then adopt more advanced process technology after the technology matures."
"I don't think so. Computer chips do not have as high a manufacturing process requirement as mobile phone chips, but the difference is actually not that big. The current top-level computer chips have basically reached a manufacturing process of 5-6nm. If we still use backward manufacturing processes, we will definitely not be able to compare with Intel, AMD, Apple and other companies in terms of chip performance. We are a latecomer in the computer field, with a brand new system and ecosystem. If our chips are still far behind them, we will have no competitiveness at all."
Hearing everyone's discussions, Xu You smiled gently.
