Global monopoly of technology

Luo Sheng did a full-body spa by the way after eating breakfast, and started working around 10 am, but the location was in the Science and Technology Complex.

Today is the weekend. As for the rest day, Luo Sheng will no longer deal with the company's affairs at this point in time. Instead, he will do his own private research and tinker with his own black technology.

Unless it's the company except for something particularly urgent.

"Master (o▽)o, the latest batch of hardware equipment from Sugon Technology will be delivered tomorrow."

Just as Luo Sheng sat down in his private office, Xiaona's voice came to her ears.

Luo Sheng nodded and said casually: "First put it in the warehouse."

Xiaona: "Master wants to upgrade Xiaona again?"

Luo Sheng sat at the desk, leaning back on the chair, holding the back of his head in his hands and closing his eyes thinking, "Yes."

Xiaona: "(⁎˃ᴗ˂⁎) The host is great, super looking forward to it, ~~"

The upgrade of supercomputers is of course very dependent on the advancement of semiconductor technology, but the more important thing is the system architecture and algorithms. Algorithms will always be the soul and core of computers or artificial intelligence.

Similarly, to achieve an instruction, the A algorithm needs 100 steps to be realized, but the B algorithm can be realized in only 10 steps, and the judgment is made on which is better.

Luo Sheng is going to forcibly upgrade his personal supercomputer this time. The hardware is definitely not available in a short time. This is too dependent on the technological development of the semiconductor industry. Now the semiconductor industry represented by Intel can only achieve 45~ 32nm process technology level.

Looking at the entire semiconductor chip industry, toothpaste factories are the first chip manufacturers to launch 32-nanometer process processors, and the mainstream process is still 45 nanometers. As for the 7-nanometer and 5-nanometer process processors in this era, they are simply sci-fi level.

The current chip technology in China is hard to describe in one word, not to mention it.

Obviously, you don't need to think about hardware upgrades for a while. You can only improve the performance of private supercomputing on the basis of existing hardware from the two paths of system architecture and algorithms.

For outsiders, this is very difficult and super difficult, but for Luo Sheng, developing a new algorithm is a very simple thing, even one person can do it, and now Xiaona can help him further improve development efficiency.

After the upgrade is completed this time, the performance of private supercomputing in the science and technology complex will increase by about ten times, from a trillion calculations to a petascale.

Called black technology.

At present, Luo Sheng's private supercomputing "Naf", named after Xiaona's official name, has a performance of 230 trillion operations per second.

In other words, Luo Sheng will reach a computing speed of at least 2,000 trillion operations per second after the upgrade. It is estimated that it will take about one and a half months to complete the upgrade plan. The hardware needs some replacement. The total budget for the major upgrade is about 120 million yuan.

By then, it will surpass the current U.S. Department of Energy's "Puma" located at Oak Ridge National Laboratory.

In the 2009 ranking of the top 500 supercomputing companies, the "Puma" currently tops the list with a calculation speed of 1,800 trillion operations per second, and is a civilian computer mainly used to simulate climate change, energy generation and other Basic science research and so on.

Of course, Luo Sheng's personal supercomputer can handle these tasks without pressure, but the fundamental purpose is to upgrade and optimize Xiaona. This is the first priority. Xiaona's upgrade and optimization bring about an improvement in overall benefits.

Has unparalleled significance.

Upgrading and optimizing Xiaona is only part of Luo Sheng’s personal work. It takes more than a month to complete this round of updates, and it is impossible for him to spend all of his time on this, otherwise it can be done in half a month. .

Luo Sheng has not officially stepped down as CEOs of the two companies. He has to participate in some major meetings, but most of the time has been spent on technology research and development.

Since entering 2009, Luo Sheng's work has basically focused on development.

Beginning in May, he broke away from the cloud computing development team of Bluestar Technology. He has left the BlueStack open source cloud platform and basically handed it over to the following technical team and Lu Qi to operate this business sector.

Luo Sheng focused on the research and development of the third-generation azure mobile phone on the azure coast. This year, the azure coast will release the latest generation of smart phones.

...

Lens Semiconductor, mobile processor division.

"The 45-nanometer process introduced a high-K insulating layer metal gate configuration. This is a milestone achievement. Intel developed this technology. These two technologies are actually designed to solve the same problem: that is, in a small size. How to ensure the effective operation of the grid."

The speaker is Xu Zhijun. He currently serves as the CEO of Lens Semiconductor, but he does not have much right to speak in the mobile processor business unit. Here is Luo Sheng who has the final decision right, and it is also a business that Lens Semiconductor has more privileges. Department, and there is an independent R&D building, which was agreed upon when the two parties jointly established the company.

But both sides have broken through the barriers to talents and patents, which is critical.

Luo Sheng was also present at this meeting, and several scientists, research scholars, and some mathematicians, etc., all gathered in this conference room to discuss.

The focus of the discussion now is the processor of the third-generation blue mobile phone. The chip processor of the new product uses a 45-nanometer process technology. This has been determined, and many technologies have been completed.

A senior research expert looked around and said: "Our R&D department thought of using metal as a gate, because metal has an effect called mirror charge, which can neutralize the dipole pair channel in the insulating layer of high-K materials. And Fermi level. In this way, you have the best of both worlds, but what these metals are, except for the western high-tech companies that master the technology, no one outside knows. This is their trade secret. ."

The insulating layer is the most critical one among all the components of the transistor, and its function is to isolate the gate and the channel.Because the gate switch channel is carried out by the electric field, and the electric field is generated by applying a certain voltage on the gate.

Ohm's law tells humans that when there is voltage, there is current.

If the current flows from the gate into the channel, what kind of switch is there to talk about?

It has long been leaked.

Another participating research scholar said: “Although silicon dioxide is good, it also has problems when the size is reduced to a certain limit. During this process, the electric field strength is constant, so from the energy band point of view, because The volatility of electrons, if the insulating layer is very narrow, there is a certain probability that electrons will have a tunneling effect and cross the energy band barrier of the insulating layer, causing leakage current."

As he said, the scholar gesticulated: "It can be imagined as passing through a wall higher than itself. The magnitude of this current is negatively related to the thickness of the insulating layer and the'barrier height' of the insulating layer. Therefore, the greater the thickness Smaller, the lower the barrier, the greater the leakage current, and the more detrimental to the transistor. On the other hand, the transistor’s switching performance and operating current all require a large insulating layer capacitance."

"It can be seen that there have been a lot of contradictions in the design goals. Whether the thickness of the insulating layer should continue to shrink. In fact, before this node, the silicon dioxide has shrunk to a thickness of less than two nanometers, which is ten. With the thickness of several atomic layers, the problem of leakage current has replaced performance problems and has become the number one enemy."

At this point, all the scholars in the conference room were quiet.

The people sitting here are the smartest group of humans, and of course they start to think of a solution when they have problems.

Human beings are very greedy, and are neither willing to give up the performance enhancement of large capacitors, nor to take the risk of leakage.

The current situation is that a material is needed, the dielectric constant is high, and the energy band barrier is also high, so that the capacitance can continue to be increased without reducing the thickness.

In other words, it is a material that protects leakage current and can improve switching performance.

In academia, in recent years, a variety of new designs have been proposed, such as tunneling transistors, negative capacitance effect transistors, carbon nanotubes, and so on.

But in fact, all these designs are basically based on four directions: materials, mechanisms, processes and structures.

At this time, Xu Zhijun glanced at Luo Sheng and then looked around and said: "What about graphene transistors? The second idea of ​​graphene as a channel is transportation, because the electron mobility of graphene is much higher than that of silicon."

As soon as this remark came out, a scholar shook his head and said: "This kind of magical material is indeed very imaginative, but the problem is that graphene transistors have not made much progress. Most of them are conceptual and theoretical stages, and graphene has a flaw. It just can’t saturate the current, but I have noticed that some people in the Western academia have said that in the future, it may be possible to control the band gap of graphene from opening to closing. It is no longer just zero band gap. In any case, graphene is a promising future Materials, but it’s too early to talk about this."

Xu Zhijun immediately looked at Luo Sheng and asked, "Mr. Luo, what do you think?"

At this time, Luo Sheng was looking at the data information and a theoretical model on the big screen of the conference. Everyone also turned their eyes on him. The latter stared at the data and function model on the screen and said:

"Listening to your discussion, I have been thinking about a question. Since electrons move in the channel, why do I have to leave such a large depletion layer under the channel? Of course I know the reason, because of the physical model This area is needed to balance the charge. But in a short-channel device, there is no need to put the depletion layer and the channel together and wait for the leakage current to flow in vain."

A group of scholars looked at each other. They had never thought about this problem.

Luo Sheng snapped his fingers, motioned to a scholar, then stared at the data model on the conference screen and said:

"Modify as I said, take this part of silicon directly to me, and replace it with an insulating layer. Under the insulating layer is the remaining silicon, so that the channel is separated from the depletion layer because the electrons come from the two poles. But the two poles and the depletion layer are separated by an insulating layer, so there is no leakage except for the channel."

Everyone's eyes lit up, as if the clouds were seeing the fog.

Everyone is among the people with the highest IQ among human beings, and Luo Sheng understood it instantly.

Xu Zhijun exhilarated and said: "Brilliant idea, Mr. Luo deserves to be a design genius, in a word, it is very important! Why didn't so many of us think of it?"

...