Tech Hammer

Jiang University, School of Economics and Management, female dormitory.

Several girls were surrounding Yang Xiaohui, looking at the problem-solving steps sent by Ning Wei.

After a long while, Xu Mo sighed: "The God of Learning is indeed a God of Learning. He doesn't look unassuming, but he is really professional when it comes to solving math problems. Unlike Xu Ruixuan, that stupid winter melon, he is the most professional when he brags!"

"Sister Mo, you're wrong to say that. A boyfriend doesn't have to be good at math. Although Mr. Xuan is not very good at answering questions, he knows how to be romantic! The flowers I gave you are beautiful, and there are little surprises from time to time. . Luo Xiang is different. Not only is he not good at studying, he is also stupid!" Lu Jiajia curled her lips and said.

"Okay, please stop saying a few words. Luo Xiang is pretty good to you. Wouldn't it be nice to make you happy every day?"

"He's just a talker, okay?" Lu Jiajia argued.

But speaking of it, the two of them are really suitable for each other. They both have cheerful personalities. When they are together, they always have endless things to talk about and are never afraid of being left alone.

"Hey, Xiaohui, what's the situation between you and Ning Wei? I think someone like Ning Wei must be focused on academics. If he is not proactive enough, you can be more proactive!" Xu Mo suddenly jumped up Say this.

Yang Xiaohui blushed instantly and said angrily: "Xu Mo, what are you talking about?"

"I'm telling you, I'd rather be with such a high-quality boy. If I miss it, I might never meet him again in this life." Xu Mo continued without any concern.

"That's right, sister Xiaohui, if I were you, I would take the initiative to pursue that Ning Wei." Lu Jiajia also supported him.

Yang Xiaohui thought for a while and said seriously: "Ning Wei is quite good, but I have also thought that being good may not be suitable. It's not like you don't know that he stayed in the laboratory for half a month, and even a few people in the dormitory were... You can't find anyone else. Do you think this state is really suitable for falling in love? And in front of him, if you are crushed by your IQ, you will really feel inferior."

"Hey, what's the point! Give yourself a chance to try first, and if it's really not suitable, break up. After all, it's fate to meet each other by this chance!" Xu Mo said disapprovingly.

"Alright, Sister Mo, let's not get involved in emotional issues. Xiaohui can handle it herself. Besides, you've covered her dormitory. Is this really good?" Yu Hanzhen, who was still single in the dormitory, said quietly. said.

"Haha, classmate Xiao Yu, don't say I didn't remind you, Liu Cong is still single! Or is he really married?"

"screw you!"

The rough-and-tumble mode was turned on in the dormitory. Yang Xiaohui looked at it with a smile, but her mind was a little dazed. Or should she really give it a try?

That night, Yang Xiaohui couldn't sleep.

…

Similarly, Ning Wei almost didn't sleep all night.

Of course, he didn't have such complicated thoughts and spent the whole night studying domestic and foreign EDA software.

I chose EDA because this software is most interested in algorithms, and algorithms happen to be Ning Wei’s current strength.

China also has its own EDA industry. Of course, compared to the three EDA giants Synopsys, Cadence and Mentor Graphics, the market share of the domestic EDA industry is pitifully weak, and the gap is not generally large.

This also includes China Jiutian, the leading company in China’s EDA industry.

According to the data, the three American EDA giants account for 95% of the global EDA industry market, and all Chinese EDA companies add up to only 1% of the market share.

If it is a high-end complex chip design, such as a CPU, it is almost zero percent.

In fact, it is not just design. An excellent EDA software will participate in it many times during the entire chip manufacturing process. Therefore, the functions of EDA can be roughly divided into four categories. The first category is manufacturability design, and It is often referred to as DFM; the second is process simulation. The difficulty in this area is that it requires massive data support;

The third is to design the interface. The main difficulty lies in providing PDK and libraries.

Especially PDK.

In Ningwei's understanding, PDK is equivalent to a direct bridge between IC design companies, foundries and EDA software vendors. A complete and excellent PDK is a set of documentation for the reaction process for foundries, while IC design companies need to use it for physical verification, which is directly related to the success or failure of tape-out.

Fourth, directly analyze the yield rate during the production process to solve the problem of improving the yield rate.

So a special EDA tool began to take shape in Ning Wei's mind.

It is called a dedicated EDA tool because although this tool covers almost all the functions of current EDA software, it cannot be used to design extremely complex general-purpose chips such as CPUs and GPUs. The main application areas are the design of specialized smart chips such as semi-customized, fully customized and brain-inspired chips.

It’s not that I don’t want to get there in one step, but it’s really too difficult. On the one hand, the workload is too huge, and on the other hand, the existing domestic foundation is relatively weak.

However, if this EDA can really be designed, it will also be unique in the EDA industry. The key lies in the project library it provides.

According to the design concept that emerged in Ning Wei's mind, he will fill many intelligent algorithms into the engineering library, so that when designing a dedicated smart chip to implement a certain function in the future, he can easily call it from the engineering library.

Money aside, if this EDA is really going to be built, Ning Wei feels that he is an expert with half a foot in the field of artificial intelligence.

During this process, some new algorithms were developed. After the entire project, he felt that it would not be a problem to win another Scientific Progress Award.

The most important thing is that as long as this EDA software is proven to be useful and popularized, he can breathe a sigh of relief. At that time, even if his brain returns to its normal state, he will still have enough foundation to establish himself.

Therefore, this software should be designed to be fool-proof. As long as the operation is simple enough, it will be easier for people to get started and it will be easier to promote. After all, the original intention of artificial intelligence development is to make lazy people lazier.

There is also a compatibility issue.

In this area, Ningwei feels the need to design a unified standard. The chips designed through this EDA software will use a unified software and hardware interface and tool chain, so that applications based on its own chips can be freely transplanted.

This requires a strict definition of the functional boundaries of the chip.

In this way, one night, Ning Wei recorded whatever he thought of. As the night passed, a new research report began to take shape. This is a big project, and if you want to complete it before graduation, you will be pressed for time.

But thinking about the researchers assigned to him by the laboratory, Ning Wei felt that it wasn't too difficult.

Of course, the premise is that the leadership pays attention to it, really treats it as a key project, and sends senior researchers instead of a few students. Ning Wei planned to find a time in the past two days to submit a report to Professor Lu and talk about the project.

The only problem is probably that carrying out this project may conflict with the postgraduate entrance examination.

But what the hell!