Wang Anran understood Guo Huaisheng's excitement very well, but Wang Anran could not make any compromises when it comes to the future development of Huaxin Technology's graphics card business unit.

Unexpectedly, Lao Guo was so angry that he proposed to the Human Resources Department to resign in the afternoon after the second-generation graphics card was cancelled and formed a resolution, which made Wang Anran unexpected.

"Director Guo, should you think about it again? I can guarantee that this resolution does not involve you personally."

Guo Huaisheng shook his head and said, "I'm tired, I just want to take a break!"

Wang Anran said: "The company can give you a long vacation, when will you adjust your condition and come back to work?"

"Forget it, Mr. Wang, I don't want the salary of a white-collar company."

After Guo Huaisheng finished speaking, he nodded to Wang Anran.

Wang Anran stood up, walked to Guo Huaisheng, shook hands, and said, "Since that's the case, everyone has their own ambitions, I won't force it anymore. When will I want to come back, Huaxin Technology welcomes you.

Of course, if you have any difficulties, you can call me directly.”

"Thanks!"

Wang Anran sincerely tried to keep him, and Wang Anran expressed his greatest respect to every scientific research talent trained by Huaxin Technology.

Some people cultivate them into seeds and fall outside to germinate and become vegetarian.

As for Guo Huaisheng, Wang Anran planned to become a technical master of Huaxin Technology.

Besides, Wang Anran felt a little sorry for what just happened.

Seeing Guo Huaisheng say thank you, he walked out of his office.

Wang Anran sat on the sofa.

I looked back at the whole thing and calmed down, and then I realized that if I changed the way of handling it, there would be different results.

The decades of technical career have allowed Wang Anran to make no progress in handling human relationships. Whenever he finds out afterwards that he has caused harm to others unconsciously.

Wang Anran shook his head and calmed down his mood. At this time, there was no time for him to reflect.

That night, Wang Anran held a meeting of the graphics card business department.

At the meeting, Wang Anran proposed the concept of alu logical computing unit for the first time.

“…We can easily find that the rendering architecture has the disadvantages of inefficiency in experimental simulations and daily processing of display chips.

Moreover, we spend a lot of algorithm optimization and circuit optimization to solve how to improve the operating efficiency of the processing unit of the rendering pipeline.

But in the face of complex application scenarios, our graphics card architecture always has inherent shortcomings…”

All of them here are technical core personnel of the graphics card business department. Wang Anran's words immediately resonated with everyone.

Yes, the processing of graphics involves modeling, rendering, rendering, and color making and mapping. Each item has a corresponding processing unit, which is specifically responsible.

The advantage of this architecture is that from modeling to rendering, full assembly line construction.

But the problem is also obvious. When the graphics modeling and rendering tasks are not input according to the actual processing capacity of the processing unit, this will inevitably cause the processing unit to continue to wait for the information feedback from the previous process after completing the work in hand.

The direct consequence is that the processing unit functions are idle, resulting in inefficiency.

This problem is almost unsolvable in rendering architecture design. The reason is very simple. The chip is dead, and the proportion of various types of processing units is also dead. Faced with a flexible and diverse processing environment, it is impossible to achieve strict processing every time.

Most of the people here have a question, what should I do?

This is what Wang Anran wants to ask them. He doesn't want to tell the answer every time, as this will make the technicians in the graphics card business department less fun.

"Three days, you have three days, and I need each of you, or your team, to report a solution idea."

After Wang Anran finished speaking, he stopped the meeting. He had already assigned homework. Now he will see how far the people in the audience can do.

A group of top students from key universities such as the University of Science and Technology, the Academy of Sciences, and the Qing University are the intellectual responsibility of the country.

In addition, Wang Anran has already prompted the answer, which is the concept of "alu logic computing unit". Wang Anran is looking forward to how far everyone's imagination can be opened.

He can do a lot of things in three days, so Wang Anran will naturally not wait stupidly. At this time, every second is very important to him. While paying close attention to the progress of the Bism project, Wang Anran turned his attention to the chip processing technology.

Good news came from ibm's chip No. 1 production line, and for the first time, the experimenters stacked two layers of transistors on a wafer.

After receiving the news, Wang Anran also came to the processing workshop as soon as possible.

"Mr. Wang..."

"Mr. Wang!"

The technicians on the scene couldn't hide their excitement. After all, they were already ahead of other manufacturers in 3D transistors and made very meaningful attempts.

Wang Anran was also infected by the anger at the scene. Under the leadership of technical director Lu Renjie, Wang Anran saw this lens in the sealing and testing workshop.

Lu Renjie introduced: "This nahdflash memory chip, with double-layer stacking in the same area, can achieve a 70% increase in capacity, while our test data shows that other technical indicators have not decreased significantly."

Wang Anran nodded. In his opinion, it is much easier to stack 3d transistors in the 1.5 micron manufacturing process than to stack 3d transistors in the subsequent 7-nanometer manufacturing process.

This is similar to the difficulty of repairing a seat clock and mechanical watches. The more precise or the higher the accuracy requirements, the greater the difficulty.

This structural principle can be clearly seen from the 3D transistor model in the laboratory. It is like a child building blocks, layer by layer online.

Of course there are two difficulties.

The first is to design. There is a logical connection between each layer of transistors within a 3D stacked chip, and local or global communication is also carried out through specific buses.

This is the new design requirements put forward in the insulating layer, semiconductor layer, metal layer and corresponding wires. It cannot be solved by laying logic circuits like the previous 2d plane.

The second one is naturally chip processing. You should know that each process of chip production has a corresponding yield rate.

Suddenly, if the process is twice as high, the yield rate will naturally drop significantly, which will lead to an increase in costs.

Price is one of the key factors in whether the product can be promoted smoothly.

The most suitable chip for 3D stacking is on the memory chip.

In 2019, Samsung, TSMC and Toshiba have basically achieved more than 90 levels.

It can be said that there is great potential in this technology.

Moreover, Wang Anran also wants to make a little money by relying on memory chips.
Chapter completed!