Riding the wind of rebirth

Hu Changfeng was so happy for a reason. Zhou Zhi had said so much before, but he couldn't dispel Hu's doubts. It was because Hu knew where the country's real weakness lay in this industry.

If this shortcoming is not addressed, then no matter how beautifully Zhouzhi describes its ecology, it will be useless.

It's like the land is owned by someone else. No matter how beautifully you build your garden, they can take it back whenever they want, and you can't do anything about it.

If breakthroughs in rare earth materials truly enable a country to overcome technological barriers and acquire copper interconnect chip manufacturing technology, then it can begin development on this basis and eventually reach or even surpass international leading levels in the chip field.

In fact, through unremitting efforts, the information industry is no longer in dire need of resources. At least in the chip industry, such as storage and industrial control, it has achieved considerable development and advantages. If it masters rare earth material interconnect technology, which is comparable to copper interconnect technology, Hu Changfeng is fully confident that it can achieve great breakthroughs in the high-speed, miniaturized, and complex aspects of various chips.

Even if this technology can be applied to memory chips that are already very mature, the impact on the world will be enormous. The multi-layer stacking technology that Huaneng Group has mastered can achieve rare earth material interconnection. For memory chips of the same volume, a single layer can immediately increase the capacity by 30%, and a thirteen-layer chip can increase the capacity by four times!
This means that a 50GB portable hard drive can be upgraded to 200GB without changing its size, weight, or power consumption! And the group's current top-of-the-line 250GB portable flash drive can even reach the terabyte level!
And the reading speed can also be increased by 30%!

Once such a breakthrough is achieved in storage, and domestically developed operating systems occupy several hundred megabytes more space than Windows and iOS, can that still be considered a drawback? It's simply a necessary space requirement to match powerful functions!

For both types of memory chips, China has now fully absorbed the flash memory technologies of Toshiba, Samsung, and Sony. Relying on Fujio Masuoka's invention and subsequent research and development of multi-layer stacking technology, China has now reached the forefront of the world.

A technological breakthrough often leads to an overall breakthrough in an industry. If the rare earth material interconnect technology is as Zhou Zhi said, then regardless of the performance of self-developed CPUs, it will be a huge boost to Huaneng's memory chip industry.

Hu Changfeng understood that Zhou Zhi wouldn't just hand over such a large profit to him for nothing, so he spread his hands and said, "Between you and me, honesty is key. So, what are your conditions?"

“There are no conditions.” Zhou Zhi smiled, even with a hint of flattery: “Besides rare earth material interconnect technology, there is another important fin field-effect transistor technology. Our group wants the ministry to sign off on it. Based on these two technologies, we will develop a non-linear computing chip with our own intellectual property rights, using GPUs to replace CPUs and become the mainstream ecosystem architecture for personal mobile information terminal devices!”

"It should at least become one of the top three mainstream architectures in China and popular architectures worldwide."

Seeing that Hu Changfeng was still hesitant, Zhou Zhi continued, "In fact, our computing cards already have this capability. What we lack is only hardware miniaturization and operating system personalization. At the algorithm design level, nonlinear computing cards can perfectly handle both linear computation simulation and nonlinear graphics rendering. At the algorithm level, we have completed the design of a new generation of GPUs."

"In other words, the chip we are developing now is a new generation of GPU that integrates old chips, peripheral electronic components, and some software algorithms. The overall size and computing power will not exceed the latest Pentium II, K6-2 and PowerPC G4. The power consumption must also be similar."

"Let's talk about this Fin Field-Effect Transistor technology first. I've heard of it before, but what exactly is it?" "Fin Field-Effect Transistor, or FinFET for short, is a new technology that is different from the traditional MOSFET structure."

“The traditional MOSFET structure is planar, and its structure can be compared to a field. The bottom layer is an insulating matrix, which is like soil. The components are on top, which are divided into source and drain, just like cornfields and wheat fields. The fields are separated by ditches.”

“Unlike farmland, there is communication between cornfields and wheat fields in integrated circuits. Electrons are like farmers who need to move between different plots of land. However, their movement is not random. Instead, they cross through bridges built across the fields. In integrated circuits, these bridges are called gates.”

"From a design theory perspective, the ideal requirement for integrated circuit design is that all farmers cross the fields via bridges. However, due to the design flaws of the planar structure of MOSFETs, many farmers can cross the trench directly from the cornfield to the wheatfield without using the bridge formed by the gate. This phenomenon is called the 'channel effect,' and these unruly farmers are called 'leakage current,' which is very detrimental to circuit control."

Professor Zhengming Hu of UC Berkeley proposed a completely new idea: since ditches have limited effectiveness in preventing farmers from crossing fields illegally, a different approach could be taken. The substrate between the fields could be raised to form walls. Previously, bridges required filling the portion of the ditch they occupied with precious materials to cross. Now, however, cheaper insulating substrate is used to raise the ditches, taking up the space that originally allowed bridges to cross them. This replaces some of the precious materials with cheaper ones and also solves the manufacturing cost problem.

"In this way, the channel of the MOSFET structure becomes a tall and thin fin protruding from the insulating substrate, with the source and drain at its two ends, and the three gates attached to its sidewalls and top. This fin-shaped structure increases the area of ​​the gate surrounding the channel, strengthens the gate's control over the channel, and can effectively alleviate the short-channel effect that occurs in planar devices, greatly improve circuit control and reduce leakage current. It can also significantly shorten the gate length of the transistor. It is precisely because of this characteristic that FinFET can improve the channel carrier mobility and effectively reduce the scattering effect of impurity ions without the need for a highly doped channel."

"It means the current control between components has been strengthened," Hu Changfeng said. "Is that what you mean?"

"Yes, this requirement was not originally necessary for chips with large-scale manufacturing processes. However, as circuits become more refined and manufacturing processes become smaller, the gate's control over the channel must become stronger. Shortening the gate length and suppressing the short-channel effect, improving circuit control, and reducing leakage current have become unavoidable and must be overcome process technologies in advanced manufacturing processes of very large-scale integrated circuits."

"In terms of current technological development, Professor Hu Zhengming's invention of the FinFET technology is the most promising for commercial application in processes below 50 nanometers."

“Aren’t we still preparing to tackle 2.5 micrometers? We’re still a long way from 50 nanometers, aren’t we?” Hu Changfeng asked.

"Theoretically speaking, once we have a 2.5-micron process production line, and then use rare earth material interconnect technology, we can achieve the production of ultra-large-scale integrated circuit chips from 2.5 microns to 25 nanometers. And once we obtain fin field-effect transistor technology, we can meet the production of ultra-large-scale integrated circuit chips from 50 nanometers to 5 nanometers."

"The two technologies complement each other; one guarantees our ability to catch up with the world, while the other guarantees our ability to surpass it." (End of Chapter)