The rise of nanometers

Golden Week in October.

The weather in the north is starting to get cooler, and the leaves are turning red and yellow. The golden autumn of the lunar calendar is also the harvest season of the year.

Full of worries, Huang Xiuyuan left the land of Qilu and headed south to Lingnan.

The scenery on both sides of the train also changed from the golden plains to lush mountains and small plains as he went south.

When the train passes the Yangtze River Bridge, the color that catches the eye is still the delicate emerald green.

After traveling for more than 20 hours, he finally returned to his hometown where he had been away for more than half a year. The changes in Shanmei were huge.

With the formation of Shanmei Semiconductor Core, a large number of semiconductor-related companies have also settled in Shanmei and Huizhou.

The Suiren Department has previously purchased a large amount of land and built a lot of commercial housing. Although major domestic real estate developers are also developing real estate in Shanmei, the Suiren Department has its own basic base.

Ziguang and Datang also have conditions attached to their entry into Shanmei, that is, they have supporting real estate projects, and they can guarantee part of the funds for semiconductor projects through the sale of real estate.

At the same time, it can also prevent employees from not having a house to live in and having their money indirectly taken away by real estate developers.

Therefore, although housing prices in Shanmei have risen to around 14,000, they have remained tepid.

Because the Suiren Department and a number of semiconductor companies have arrangements for housing allocation, semiconductor practitioners as a high-income group have been retained by the Suiren Department and a number of semiconductor companies.

Without the support of high-income groups, market demand will naturally appear weak, and it will be useless for real estate speculators to speculate all day long.

Back in Shanmei, Huang Xiuyuan had lunch with his parents and then hurried to the semiconductor base in Haifeng County.

When they heard that he was back, Lu Xuedong, Zhang Jingjian and others welcomed him at the gate of the base.

As soon as they got off the minibus, they saw Lu Xuedong, Zhang Jingjian and others who were much thinner. He asked with concern: "Xuedong, you need to get more rest. Your body is the capital of revolution. And Lao Zhang, you too."

"Thank you, Chairman, for your concern." Zhang Jingjian turned sideways and moved out of the way.

Lu Xuedong, on the other hand, didn't pay much attention and changed the topic: "Xiuyuan, you came back just in time. Today happens to be the second filming."

"Which workshop?"

"Factory Building No. 5." Lu Xuedong said and led the way.

Along the way, many employees greeted him. The semiconductor base was very busy, similar to what he experienced at the Texas Semiconductor Base.

Among the faces he saw, apart from the tired expressions, the employees' eyes were full of burning eyes, which was a kind of confident light.

This is also a phenomenon among researchers from the Suiren Department, which is different from other companies in the country. In terms of spirit, people from the Suiren Department are more confident and full of fighting spirit.

Come to Factory Building No. 5.

One of the chip production lines equipped with ZY-2 is used for the tape-out of Fuxi chips.

When Huang Xiuyuan put on protective clothing and entered the testing room, more than 20 large-scale films had been taped out on the production line.

The large wafers that had not undergone deep processing were cut into semi-finished chips by a laser cutting machine. Huang Xiuyuan picked up a semi-finished chip.

The chip is very small, smaller than a finger.

Lu Xuedong introduced in a low voice from the side: "The Fuxi chip integrates 1.06 billion transistors and performed well during the preliminary test operation."

"Where are the problems detected?" Huang Xiuyuan put down the semi-finished chip.

Lu Xuedong took out a tablet and called up the relevant test situation, which listed hundreds of large and small questions.

He was not surprised by this. If it was his first time to design a chip, or such a difficult CPU chip, and he could get it right in one step, it would be absolutely impossible.

Even for mature designs, some unexpected problems may occur due to processing errors during the actual tape-out process.

Either a slight vibration, a dust particle, or the aging of the equipment may cause the results to differ.

The textile method is better. If it is photolithography, not only the processing environment requirements are very high, but the processing cost is also high.

In terms of tape-out, the single tape-out cost of the textile method is only about 200,000 to 400,000, while the photolithography method does not have tens of millions, so there is no way to do it.

Low tape-out cost is also one of the advantages of the new chip system.

Huawei, ZTE, Ziguang and Datang basically alternate design and tape-out. For example, Ziguang's high-end storage chips were tape-out five times this month alone.

If it is photolithography, based on Ziguang's tape-out frequency, when their chip design matures, the tape-out cost is estimated to be in the billions.

It can be taped out at low cost, giving semiconductor companies more opportunities to tape out their designs. This is the advantage of low-cost trial and error.

In this way, many rookie semiconductor designers have grown up rapidly.

Anyway, we just work hard to build it. If it doesn't work once, we tape it out 10 times or 100 times.

After reading the previous dozen questions, Huang Xiuyuan saw the problem with the Fuxi chip, mainly because of the first design. It was very difficult to ensure that 1.06 billion crystals could perform their respective tasks without interfering with each other.

"One-fifth of the transistors have failed to varying degrees due to mutual interference. Sure enough, they are still immature." Huang Xiuyuan shook his head helplessly.

Lu Xuedong also looked embarrassed.

"Chairman, we didn't expect it to be so serious. In the supercomputer simulation calculation, no problem was detected." Zhang Jingjian explained.

While Huang Xiuyuan continued to look at the details of the problem on the tablet, he comforted: "I don't mean to blame you. I expected this situation to occur."

There are no problems with supercomputer simulation calculations, but after actual manufacturing, there are a lot of problems. This is actually caused by insufficient database accumulation of industrial software.

After all, there are obvious differences between textile nanowire chips and photolithography chips.

In the previous chip design software, a lot of data were based on photolithography chips, which naturally led to inacclimatization.

In this case, we can only rely on designers to accumulate and improve bit by bit in research and development. In a short period of time, we can only rely on burning money and trial and error.

The semiconductor industry is indeed not something that ordinary people can afford. The technical threshold is high and it requires a huge investment, which may not necessarily be profitable.

After Huang Xiuyuan comforted them, he began to retrieve the future nanowire chip design information in his mind.

"Look at this place, this knot should be thicker..."

Under his guidance, Lu Xuedong and others took notes while listening. After listening to Huang Xiuyuan's guidance, they felt suddenly enlightened.

It took three days for everyone to discuss the design issues of the Fuxi chip. Every small problem was dissected one by one and then extended. Under Huang Xiuyuan's in-depth explanation, they avoided many detours.

The test results of the second tapeout also came out on the third day.

After getting this test data, everyone continued to discuss how to optimize and improve.

Back and forth, Huang Xiuyuan worked in the Fuxi chip project team for twenty days. After seeing the newly improved Fuxi chip and preparing for the third tape-out, he left for Huizhou next door.

Because the joint GPU project team saw so many problems with Fuxi chips in Huawei's Huizhou semiconductor base, Huang Xiuyuan was also worried about the design of joint GPU.

Thank you for your support (ω｀). I’ll make up for yesterday’s fifth update!