Rewrite the Technological Landscape

Chapter 217

On June 15th, or rather recently, Meng Qian was dealing with another matter in Hangzhou. With the progress of the acquisition of the five companies in Taiwan, Meng Qian immediately recruited a group of people to work in the company.

The design progress of CMOS image sensors is going to be accelerated. For the forward design of a chip, the first step is to determine the plan, then the algorithm engineer to get the architecture, and the next step is the ASIC engineer to get the translation work. Now in mainland China, both algorithm engineers and ASIC engineers are is extremely scarce.

Therefore, Meng Qian has always been interested in training, and gradually cultivated excellent algorithm engineers and ASIC engineers from various development departments of the company. Of course, the longer-term goal is to train architects. In the next few years, excellent architects will be very valuable. It is also very rare in the country.

Just like the algorithm architecture of the CMOS image sensor this time, Meng Qian gave a lot of space to the people in the company's system development department, but now external pressure forced Meng Qian to speed up the company's semiconductor development, and quickly recruited a group of experienced people. Continue to the next work.

For a chip design, the most time-consuming part is the initial plan formulation. Meng Qian has completed this part, and now the company has completed the structure, and the translation has almost been done. The rest of the work will not cost much. How much time is needed, what is needed more is professionalism and meticulousness.

For example, the next step to complete the translation is simulation verification. This is the most important part of all links after the entire chip design plan is determined. It will not take much time. According to the complexity of the chip design, the efficiency of the software, and the advanced computer equipment Depending on the influencing factors such as degree, the simulation verification of some low-end chips only takes tens of minutes, and the simulation verification time of high-end chips generally does not exceed half a month.

Of course, although the software technology and computer hardware were almost inferior in 2002, the chips at this time were not as complicated as later generations. For simulation verification of CMOS image sensor chips, a few days are enough.

As for the simulation software this time, Meng Qian chose to use his own simulation software without hesitation.

The company's industrial software has been laid out for so long. If you dare not use your own industrial software but use other people's products at this time, it will be a blow to the confidence of the company's employees.

If you can't trust yourself, do you expect others to trust you?

Of course, there is also a particularly important thing involved here, that is, the process file, which is the secret of the chip manufacturer and can only be obtained by the chip design company that cooperates with the fab. Another point is that any chip design from a The process must be determined at the beginning, and this must also be communicated with the cooperative chip manufacturer in advance.

Therefore, if a chip design company without any cooperative chip manufacturers suddenly comes up with a chip design, and then goes to the manufacturer to produce it, this kind of thing is basically impossible to happen in reality. This is also the reason why Meng Qian needs to cooperate with Sony. The only chip manufacturer he has a cooperative relationship with is Huaxin, but Huaxin's level is known to everyone.

After the simulation verification, it is synthesis, timing, power consumption analysis, automatic layout and routing, etc., and finally the merged layout is obtained.

But after all, the design is done, and the rest of the work is done very handily by the employees of the Taiwan province acquired. And under the request of Meng Qian, everyone works overtime every day to catch up. Although the employees in Taiwan province were a bit rebellious at first, But when they saw the company's overtime pay, overtime benefits, and employee benefits, they all had a blast.

Of course, there are always those who can't stand the intensity of the work, some resign directly, and some curse while doing it.

The pace of chip design is accelerating. On July 10, it took nearly three months. Of course, without counting Meng Qian's design time, Dafeng Technology finally completed the company's first chip design.

That day, Meng Qian went directly to Sony, but this time, Idei received Meng Qian in person. This subtle approach also caught Meng Qian's attention.

"Mr. Meng, long time no see." As soon as we met, Dejing showed enthusiasm.

Meng Qian also responded with enthusiasm, the two chatted for a while, and Idei went to the factory with Meng Qian.

"How long does it take to tape out?" Meng Qian asked the person in charge of the factory connection.

"We have basically achieved full automation now, and it doesn't take much time to tap out, mainly because the later testing takes a little time." The reason why the other party mentioned fully automatic is because chip manufacturers in the 1980s had many manual positions, and in the 1990s Only at the beginning is it slowly moving towards full automation.

Now in 2002, it is something to be proud of being able to realize the concept of full automation.

"How long will the test take?"

"At least ten days."

Meng Qian nodded and didn't say much. This matter is beyond his control. The test time in this era is generally longer than that in later generations. There is no way, it is also a technical limitation. Fortunately, the cost of taping out is much cheaper than that of later generations, and this thing is related to craftsmanship.

The high-end chips with 7nm process cost hundreds of millions of dollars at a time, and the Kirin 990 cost 200 million at a time. And Meng Qian's tape-out cost budget this time is only a few million.

But what is even more frightening about making chips is the failure of tape out. Once the tape out fails, the money spent will be spent, and you have to go home to redesign, tape out again, and pay the money again. Just like Xiaomi’s Pengpai, it is said that the tape-out failed many times.

So really not many companies dare to touch high-end chips, and they really can't afford to spend money.

"Mr. Idei, if the tape-out goes well this time, when do you think the product will be released?" Meng Qian asked Idei for advice after finishing the factory.

Idei thought about it for a while, "In fact, we will have a press conference in early August. If the taping test goes well, we can consider arranging it together."

"Early August? That's less than a month away. If you count the tape-out test time, it's impossible to prepare for mass production, right?"

"It doesn't matter whether it is mass-produced or not, as long as the product is sure that there is no problem, no one stipulates that the things released at the press conference must be sold immediately." Idei made a careless expression, "Now companies are starting to play with the future, first What is the release and mass production?"

Meng Qian smiled. It is true that as Idei said, technology companies in later generations have begun to play with the concept of the future. The products have not yet been produced, or even just in the research and development stage, and they have started to brag at the press conference.

"Of course, it mainly depends on the test report of the product. After all, you have adopted a brand-new design, which involves more things and is more complicated." Idei left a way out for himself without knowing it. Meng Qian didn't even hear this small trick, "Mr. Meng may not know that there will be a melee in the field of CMOS image sensors in August."

"Melee?" Meng Qian was really not very clear about this matter.

"The fuse of this matter is Kodak."

Idei briefly talked about the situation with Meng Qian. To put it bluntly, it is now in the era of replacement of CMOS image sensors and CCD image sensors, but at the moment CCD is still the mainstream, and CMOS image sensors are still in a state of momentum.

But companies like Micron and Sony have already seen the potential of CMOS image sensors, so everyone is secretly making efforts in this field.

Originally, everyone engaged in research and development independently, but at this time, Kodak jumped out of the memory of a generation.

As the originator of the camera field, Kodak became increasingly unable to keep up with the trend because of its long-term entanglement in the film era, and finally went bankrupt. However, at this moment in 2002, Kodak had not yet come to an end, but it had been going downhill, and its industry status was still there.

At this time, Kodak suddenly announced that it has developed a high-performance CMOS image sensor, which will beat the CCD image sensor.

This is like Nokia suddenly saying that I have developed a better mobile phone when it started to go downhill but has not yet completely lost. Its influence is very large, not only the influence of the industry, but more importantly, the influence of consumers. Kodak is a symbol of an era. Even after the bankruptcy, there are still many photographers who believe in Kodak.

Then, all companies engaged in CMOS image sensors had to jump out, because in the field of technology, a slow step may slow down a lifetime. The reason why Huawei and Samsung are so tight in the field.

"Kodak, Fuji, Micron, Sony, Samsung, Nikon, Canon... This is really a fight between gods."

The translator couldn't help being stunned for a moment, "What does it mean that the gods fight?"

"The amount is a gathering of masters."

Idei smiled after listening to the translation, "So, if Mr. Meng's new design can greatly improve product performance after passing the test, of course we can't miss this opportunity. In this melee, everyone will spend their money to compete. We can't What is mass production, someone will definitely talk about things that are still under development.”

After understanding the situation, Meng Qian didn't make any further comments, "Then let's look forward to the results of the tape-out test."

It took a long time to wait. Meng Qian did not stay in Neon Country. The next day, he took time to visit Ueno of Daiwa Housing Industry Co., Ltd., and met a few entrepreneurs in Neon Country, and then returned home.

After waiting for 15 days, Meng Qian received an email from Sony in the afternoon of 15 days, and the test data and test results were attached in detail in the email.

Just after reading the email, Idei called Meng Qian as if he was stuck, and the voices of Idei and the interpreter were on the other end of the phone, "Have you read the email, Mr. Meng?"

"I've seen it, and Mr. Idei will take care of it personally."

"You are welcome, Mr. Meng, but this test result is a bit regrettable." Idei suddenly changed the subject, "From the results, the sensor itself should be fine, but there is a problem with how to combine it with our products. For insurance reasons , At this press conference, we temporarily decided not to release Mr. Meng's sensor.

After the risks are resolved, we will release the next new product together. "

Hearing this, Meng Qian couldn't help narrowing his eyes.