After listening to Tang Shengwu's introduction.

Chen Jin nodded.

Suddenly, I remembered another thing.

He asked Ling Jundong, who was sitting next to his left hand: "Old Ling, did Intel release a desktop CPU product last month, which is said to surpass our 'Qinglong second generation', is there such a thing?" ”

Some time ago, he vaguely remembered that under his personal Weibo, there were a lot of provocative comments, and many of them were written in English.

He didn't speak English very well, but when he noticed words like 'fark' in it, he realized of course that it wasn't a good thing to say.

"Yes, there is such a thing."

Ling Jundong nodded and said, "The performance of the Qinglong second-generation CPU was surpassed by Intel's iX-10000 CPU, and their counterattack was very fast. ”

"However, the method they adopted was very extreme, they doubled the area of the chip wafer, increased the working frequency to 8GHZ, and finally surpassed our Qinglong second generation in terms of running scores, which can be regarded as a game."

"But it doesn't make much sense, the maximum power consumption of this CPU is more than 400 watts, it can be said to be a fire dragon, hot can fry eggs on it, its commercial value is not large, just to prove that they have not lost, this product can hardly be sold on the market."

Intel has only won back face, and the market is still losing ground.

"8GHZ WORKING FREQUENCY, THE CPU MUST NOT SMOKE DIRECTLY?"

Chen Jin frowned and asked, the working frequency that silicon chips can withstand, 5GHZ is basically the upper limit, how did they achieve 8GHZ without burning.

"They used the second-generation graphene heat dissipation technology, and the heat dissipation capacity has been greatly improved, and if they use liquid nitrogen cooling, they can even overclock to 10GHZ, but they can still quickly remove heat and make the CPU work normally."

"From this point of view, Intel's product does have some advantages, our Suzaku second-generation CPU, domestic DIY players are only overclocked to 6.8GHZ, and we still have some places to continue to catch up."

Ling Jundong said.

In addition, Intel intends to conquer the 7nm process CPU within 18 months, and the performance will be doubled by then.

and carried out the research and development of a new architecture, and strive to come up with a more advanced architecture than X86 within two years, to the extent that it is comparable to the "Weaver Girl Architecture".

In short, in order to meet the challenges of Xinghai Technology, Intel has mobilized with all its might, completely giving up the "squeezing toothpaste strategy", but throwing out a toothpaste every year, and Xinghai Technology must be suppressed!

However, in Chen Jin's view, Intel's actions are just a dying struggle.

Xinghai Technology will not dance with it and march to 7 nanometers.

Instead, they drove straight into the other track, leaving Intel behind forever.

Because there is a CPU made of material, which has a much brighter future than a CPU made of silicon crystals!

Graphene CPU!

Silicon chips have come to an end, and it's the real future.

In addition to the materials research center, there is a graphene CPU R&D team in the CPU research institute of Xinghai Technology, which has been established for more than two years, and Chen Jin has also invested a lot of funds and resources to provide a lot of high-value information.

In addition, the material research institute has thoroughly conquered the production technology of graphene.

The graphene CPU he was looking forward to had finally come up with some laboratory products recently, but Chen Jin did not know the specific performance of the product in detail.

So Liang Ying, vice president, contacted the research center, and in the video conference, Chen Jin consulted Rong Weizhi, the person in charge of the project, with a few questions.

"Team Leader Rong, you said that you have made several laboratory products, how are the performance of these products?"

Rong Weizhi is in his early thirties, very young, and is a top student in a famous school, with a rather arrogant personality, and is a 'perfectionist', not only demanding of himself, but also of others, so he is not very gregarious, and many people complain behind him, calling him 'extremely high IQ and zero EQ'.

Fortunately, it does not delay the progress of the project, and Chen Jin does not particularly care.

"No, the product performance is not good, it's all garbage."

Rong Weizhi shook his head: "We have determined the technical route of graphene CPU very early, there is no big problem, what we are doing wholeheartedly is to try our best to reduce the volume of graphene transistors." ”

"In August last year, we succeeded in producing the first graphene transistor, which is 100 microns in size, which is very far from the commercial standard."

"We kept improving, reducing the size of the transistors, and after a month, we were able to reduce the size to 10 microns."

"And we continued to improve, so that the transistor size continued to shrink, compressed to 1 micron, 600 nanometer, 500 nanometer, 400 nanometer, 300 nanometer...... Recently, we have been able to shrink the size by 100 nanometers every month. ”

"In August this year, we finally achieved the size of the graphene transistor to the level of 100 nanometers."

"But then it suddenly becomes more difficult, these two months have not achieved significant results, the size of 100 nanometers is absolutely not good, completely does not meet the commercial standards, must be compressed to 50 nanometers, 30 nanometers below, in order to completely replace the traditional silicon CPU possibility."

Rong Weizhi shook his head again and again, he was still extremely dissatisfied with the progress made so far.

100 nm?

Chen Jin's heart moved, and he asked: "After you compress the process of graphene CPU to 100 nanometers, how much can it work at a frequency?" ”

“100GHZ。”

Rong Weizhi replied: "It is easy to achieve this working frequency, the heat generation of graphene transistors is only one thousandth of that of traditional silicon transistors, and the thermal conductivity far exceeds that of silicon, theoretically, the working frequency of graphene CPU can be increased to 1THZ, but the heat is still smaller." ”

Chen Jin was quite surprised, 1THZ, this working frequency, can be said to be extremely exaggerated.

This means that the efficiency of one graphene transistor is equal to 1000 silicon transistors.

In the same area, the number of transistors in a 10nm process chip is about 100 times that of a 100nm process chip.

But in terms of operating frequency, graphene transistors can be thousands of times higher.

On the whole, the graphene CPU with 100nm process cannot be said to have no commercial value at all, but the commercial value is very high.

FOR EXAMPLE, THE PROCESSOR USED IN SUPERCOMPUTERS, EVEN IF THE PROCESS IS RELATIVELY BACKWARD, IF THE OPERATING FREQUENCY IS SET TO 100GHZ, THE COMPUTING PERFORMANCE IS NOT MUCH DIFFERENT, BUT THE ELECTRICITY COST IS SAVED BY AT LEAST 90%.

Based on the annual electricity cost of more than 300 million yuan for an Ai-class supercomputer, after using a graphene CPU, the annual electricity cost saved is more than 270 million.

The electricity cost saved is almost back to the cost of buying a CPU.

Due to Rong Weizhi's perfectionist character, the graphene CPU that can be commercialized has been delayed for two months.

Chen Jin immediately criticized him: "The 100nm process is enough, why did you tell us now?" Dozens of days were delayed in vain! ”

"I'm sorry Mr. Chen, it's my fault, I didn't fully realize its value."

Rong Weizhi realized his mistake and grabbed his head to apologize.

Chen Jin shook his head and did not continue to criticize him, but gave two instructions: one is to continue to explore graphene CPUs with smaller processes.

Second, the 100nm process graphene CPU will be commercialized immediately for the development of supercomputing with lower power consumption!

After all, supercomputing is too power-consuming, and Xinghai Technology has a lot of large-scale servers and supercomputers, of which the number of supercomputers has exceeded 30, reaching the standard of opening a supercomputing Internet café.

Naturally, there is an incomparably exaggerated electricity bill.

At least three or four billion yuan of electricity bills per year, even a local tyrant like Chen Jin can't completely ignore it...... I thought about how much it would be nice to save some of my electricity bills.

Therefore, the use of graphene CPU can be said to be urgent.