1991 Starting from the core

Su Yuanshan glanced at Chen Jing, who gave him an affirmative look. But... Chen Jing didn't know that what Su Yuanshan was thinking about at this time was not the acquisition of the SIS team, but something else.

In the past eight years, Yuanxin has risen. At the same time, the rise of Yuanxin also means the decline of other parts of Asia.

The overall market is only that big. When the rise speed and size of an enterprise exceeds the increase in the market and industry itself due to economic development and technological progress, the enterprise will inevitably plunder or encroach The market share, resources, and opportunities of other similar enterprises.

——If the theory is placed in the rise and competition of countries, it is also valid.

Thucydides' trap, zero-sum game, and winner-take-all all come from this.

//Note: The domestic view is critical of this statement, because it is the Western view and attitude, and the domestic view is win-win...emmmm, everyone understands the rhetoric of the harmless white rabbit.

Therefore, with the rise of Yuanxin and Huajing's follow-up, the most intuitive impact on the island is that... their semiconductor industry has failed to rise as it did historically.

According to the latest semiconductor industry revenue statistics, Deyuan's revenue has successfully entered the top five in the global semiconductor industry, becoming the first wafer fab to enter the top five purely relying on foundry production. And Huajing also relied on avoiding the semiconductor winter in China, and thus directly touched the tail of the top ten from ranking fifteen.

As for the island, two fabs that were supposed to take advantage of the 1997 Asian financial crisis lost their last chance. Now UMC has fallen to ninth place, and TSMC has fallen behind fifteenth place.

The ripple effects don't stop there.

With the active expansion of Hongyuan Foundry and the role model of Meijie as the big brother, the Special Zone has become a good place for many board and card factories on the island.

In addition, last year, a group of Chinese survivors from Java gathered in Fujian Province and finally chose to invest with the government to build a large-scale technology industrial park - mainly producing computer boards and consumer electronics.

This science and technology industrial park is said to have attracted several PC card-related companies on the island to move in.

It is not that the island has no control over this, but today the island no longer has the absolute, near-monopoly advantage in the PC field as it did ten years ago.

Are you kidding me? Deyuan is one of the world's top and most technologically advanced professional foundries, and a chip design company as powerful as Yuanxin is located in mainland China.

Not to mention, the mainland's photolithography machines have taken the lead in realizing dual workpiece stage technology, coupled with preferential policies and talent cost advantages... How can the island compete?

Therefore, any semiconductor company with a little bit of savvy will clearly realize that there is only a dead end if they continue to stay on the island.

Therefore, just after October, the island quietly lifted the restrictions on wafer factories on the mainland. At first, the island only allowed wafer factories to invest and build wafer factories on the mainland that were one to two generations behind, thus ensuring The island's technological advantage over the mainland.

With the rise of Deyuan, this restriction has become a policy that forces the island's wafer fabs to stay. Just kidding, as direct competitors, UMC and TSMC themselves lag behind Deyuan in terms of manufacturing processes. It is only slightly ahead of Huajing's first generation process. If you build a factory in the mainland and you still have to abide by the regulations and build a wafer factory that is a generation behind the island, then I'm afraid it will be a waste of money...

Therefore, just last month, UMC and TSMC were also actively negotiating with the mainland government to prepare to open a second battlefield - we can't just watch the island's semiconductor industry die out, right?

…

As for Chen Jing, as a girl who was born on the island, Su Yuanshan believed that she obviously had feelings for the island.

Okay, SIS is SIS. Su Yuanshan nodded and quickly agreed.

Chen Jing also nodded and quickly jotted down a few notes in her notebook - the executive board was just proposing ideas, and the plan had to be finalized before it was finalized.

But let me say one more thing. Su Yuanshan looked at Mr. Xi again: Senior brother, you can't stay out of this matter, just say that you are the boss, and then wait for the chip to come out in a year or two - the heterogeneous parallel computing solution ...Well, how about you take a look, sort out your thoughts, and make a report, and I'll give it to Academician Su.

Academician Su?

...Uh, my dad. Su Yuanshan coughed and chuckled: Speaking of which, senior brother, when should you be evaluated as an academician?

I don't ask for that. Xi Xiaoding waved his hand and glanced at Su Yuanshan with disdain - this guy actually called his father such a bureaucrat: You mean, you want to push the Electrical Engineering Department to move towards heterogeneous supercomputing development?

Academician Su Xinghe has been working on supercomputing projects before he became the president, and he even plans to work in the field of supercomputing until he is old. Now he is a principal, but he still holds the title of supercomputing project leader - after all, this is his hard work.

Well. I thought about it just now. Currently, Dianke's supercomputing is still taking the isomorphic route, but heterogeneity is definitely the future direction. If Dianke can take a step early, it will have a say in heterogeneous supercomputing in the future. It’s one more point.”

In addition, I agree with the idea of ​​3DFX's multiple arithmetic logic units and threads that you mentioned before. After all, compared to the processor, it is more professional in terms of computing. At the same time, it also means that it is necessary to achieve those needs. The best processing core for computationally intensive algorithms.”

Xi Xiaoding nodded with a smile: That's the reason. After getting Xinghai's GPU design, we simply simulated it and found that it is much faster than running a convolutional neural network on a processor.

So, you can't stay out of it. Su Yuanshan said with a smile again: When we build GPUs, we can't just focus on the existing function of 3G graphics acceleration, but also look to the further future.

After speaking, Su Yuanshan closed the notebook, indicating that the topic was over, and then looked at the others again.

Any other ideas?

Chen Jing glanced around and saw that no one had the intention to speak. She endured it for a while and said softly: Let me speak.

Su Yuanshan smiled at her and nodded.

A technical field and an enterprise field are exactly in line with the structure of Yuanxin.

“I believe that this year, while we maintain expansion, consolidate results, and open up new areas, we must also focus on the exchange and cultivation of talents.”

As she spoke, Chen Jing looked around the crowd and said slowly: I wonder if you have realized that Yuanxin's current atmosphere and structure are not very suitable for internationalization.

After her words fell, everyone became thoughtful. Su Yuanshan, in particular, looked at her hesitantly.