Wei Haoguang picked up the document and looked at it, and the more he looked at it, the more frowning he became...... After a while, he said, "Su Han! What do you write about this development plan...... It's a little too out of the mainstream! A lot of the features above, a bit impractical. Especially if you plan to completely break away from the complex instruction set and build an independent reduced instruction set chip system. Will there be a market for this? ”

"There is definitely a market. In fact, the miniaturization and integration of computing chips have become a trend.

I think that in the future, the CPU field will eventually form a two-strong split situation.

The first is based on the complex instruction set represented by X86.

The second type is a micro-miniature chip designed based on a reduced instruction set.

Chips with complex instruction sets have little to no consideration for the cost of power consumption. Because desktop computers can supply a stable power supply, there is no need to worry about power supply.

But as society progresses, I think there will be a blowout in the use of CPUs on mobile. It seems that mobile phones nowadays are not a rarity. There is even a new type of mobile phone that can transmit more digital information, so with the development of the times, the amount of information transmitted by the mobile terminal will only increase.

When the mobile terminal becomes more convenient and fast, and even everyone in the world has a mobile phone, the demand for mobile chips will eventually surpass the desktop and become the most important chip development area in the world.

Then we have to think about a long-term development plan.

One thing must be admitted, it is almost impossible and unnecessary for us to try to catch up with the Mi Federation in the complex instruction set now. However, it is still possible to lead the world in the field of mobile chip design, which has not yet developed. This is also known as overtaking in corners.

In my opinion, the reduced instruction set chip of the future should have the following characteristics.

One is low power consumption. Because the mobile terminal needs to be powered off frequently. If the power supply is stable, it must consume a lot of power, so it is a chicken rib product. Therefore, the chip must save power, for example, it can dynamically adjust the voltage or something, reduce power consumption, and reduce heat.

The second is multi-core and large cache. The more cores integrated by the chip, the stronger the multi-task execution and processing ability, and the lower the CPU occupancy. There is no need to adjust the voltage frequently, so the power consumption is less and the heat generation is lower. The cache is larger, the time efficiency is high, and there is no need to waste efficiency and fragment computing, which will naturally save more power.

3. Cloud computing. This is a computational concept that I proposed. As long as the chip is properly designed, the distributed computing amount can be connected to parallel computing. Such individual calculations are not fought in isolation, but can also be turned into a whole. At that time, it will not only reduce the computing cost, but also be collaboratively deployed and dynamically adjusted, which I think should have a strong application prospect. The premise is that network technology can accelerate progress. At least a small scale of cloud computing can still be completed.

Fourth, malleability. I have two ideas for this one, and if you want to improve scalability, you have to be modular. If we want to build a closed instruction set, it's okay for a short time. However, after a long time, as the iteration of the version increases, the architecture will definitely become cumbersome and complex, and the compatibility will be poor, and the threshold will increase itself when the time comes, which is not suitable for large-scale promotion.

However, if you create a relatively open-source modular architecture and only charge a small license fee, customers can customize the chip according to their own needs to adapt to a variety of instruction sets, which will have a lower threshold and be more conducive to promotion.

Of course, it can also be completely open source without charging any patent fees. An international platform has been created.

In this way, it is easier to quickly occupy the market, but the later income is limited. However, you can consider establishing an instruction set chip design company, if you do not pay the license fee, you have to pay the design drawings, and then the company takes technical support or selling drawings as the starting point for profits, which can also further increase profits.

If you don't make money as the first goal, but take national security, technological security, software and hardware ecology, technology controllability, and internationalization as the goal. Open source is also possible. ”

Wei Haoguang took a deep breath when he heard this, and he had to say that Su Han's statement helped him open the door to a new era, but he didn't expect the other party to think so much and think so far, so he was not convinced.

The key is that Su Han's suggestion allows him to continue to iterate on the Risc design ideas that he has worked so hard to research.

That's the most important thing.

At present, there is a serious academic waste in China, and after researching a topic and achieving its goal, no one will use it, and it is really distressing enough to throw it away.

Wei Haoguang thought for a while and said, "This matter! Let's discuss. There has to be someone! You wait. After saying that, he got up and picked up the phone and dialed it directly.

……

It didn't take long.

There was a knock on the door.

Wei Haoguang got up to open the door and led a man into the room.

Wei Haoguang said: "Su Han! Tell me about you. This is the dean of the School of Microelectronics of our school, Jiang Binglu, Dean Jiang. Binglu! This is my student, Su Han. ”

"Hello Dean Jiang." Su Han hurriedly got up to say hello.

Jiang Binglu smiled: "Su Han, right." Your name is thunderous right now. We are the first in our country to design a computer assembly language with independent intellectual property rights. Your current name is already known to everyone in the school. ”

Su Han smiled and didn't say anything.

"Binglu! Sit! I have important things to say. Wei Haoguang let Jiang Binglu on the sofa, then took out the documents prepared by Su Han and handed them to the other party.

Jiang Binglu looked at it for a while...... His face also changed from relaxed to solemn.

As an expert in the microelectronics industry, he certainly knows what is written on it. Although he didn't look as far as Su Han, he could also sensitively perceive the design ideas of this computer instruction set and the great influence it might have in the future.

After reading it, Jiang Binglu said: "Su Han! What is written on this is your thoughts? ”

Su Han nodded and said, "It's all my thoughts, I discussed this issue with the teacher before......" Speaking of this, Su Han repeated what he had said to Wei Haoguang before.

Jiang Binglu kept nodding his head when he heard it...... I have to say that Su Han's words are word for word, and every word has been said to his heart.

Jiang Binglu said: "Su Han! You're right. Our country should have a fully controllable computer instruction set and architecture. This is not only related to national security, but also related to the overall future development of the entire country's computer industry.

But now the problem is that the technology of our country is already much behind the West.