Earth simulator supercomputing is 35 trillion.

The performance of the two sides is simply a difference between heaven and earth.

At present, the international supercomputer is either the Mi Federation or the Long Island country.

Other countries have basically zero presence in supercomputing.

So no one takes Huaxia's 13 billion supercomputing seriously.

However, when Longhan Technology officially announced the various data of Shenlong Supercomputer, it was a global sensation.

Shenlong Supercomputer has a total of 15,000 nodes.

Each node has two Hualong H3 processors. There are also two graphics card chips developed by Longhan Technology, which also use the eight-core Cailong C1.

In fact, the Hanlin Academy has long developed a graphics card chip, but Su Han has not been introduced to the market.

This will be directly used in the dragon supercomputer.

Shenlong Supercomputer has a total of 360,000 computing cores.

Computing power has reached a staggering quadrillion per second.

This data directly shocked the world.

Because the computing speed of the Shenlong supercomputer is actually twenty-eight times faster than that of the earth simulator.

That's an exaggeration, isn't it?

Almost all major news is talking about the new king of the supercomputing world: Shenlong supercomputer.

Everyone knows such awesome data.

I'm afraid that no one will be able to break this record in a short period of time.

This made CBM, which was developing supercomputing at the same time, a little dumbfounded.

Because the new generation of supercomputing computing power they are developing is only 70 trillion times.

I thought this result would definitely crush the Earth Simulator!

But in front of the 100 trillion supercalculations of the dragon is nothing at all.

……

"Huaxia Supercomputer: Shenlong. The new king of the supercomputing world. Major media outlets have also reported on it.

……

Of course, the top brass of the Mi Federation was a little unhappy.

Because in the field of supercomputing, it has always been the world of the Mi Federation.

Although there are occasional Long Island countries that come to cause trouble, the overall impact is still not significant.

But the Chinese people have also joined the competition for supercomputing.

This is definitely not good news.

Huaxia is too large and strong in the industrial field, and it is definitely a heavyweight opponent.

Huaxia Kingdom now has such a terrifying supercomputer, which will only make them more powerful in the future, which is definitely extremely unfavorable to the Mi Federation.

Mi Federation allows CBM to accelerate the research and development of supercomputing.

We have to get back the number one position in the world.

……

But this is the obvious layman's guidance to the insider.

The iteration of supercomputing requires several years of research and development, and finally the product can be formed.

The two generations of supercomputing that CBM is developing are not the opponent of Shenlong supercomputing in terms of planned performance.

In other words, I want to develop a more awesome supercomputer.

You have to start from scratch.

The key problem is that the performance of CBM's self-developed chips is far inferior to that of Hualong H3.

You can't let them use Intel or Am chips!

That would definitely be a slap in the face for CBM.

Now that I think about it, I'm afraid that in the supercomputing market, Shenlong Supercomputer will be rampant and domineering for a while.

……

Su Han didn't care about these things.

What he is more concerned with is that his artificial intelligence program will be integrated with Shenlong Supercomputer as soon as possible.

However, Su Han also threw an olive branch to Cui Xin, hoping that he could join his company and continue to develop the next generation of supercomputing.

Cui Xin, as the head of the supercomputing department of UCAS, is regarded as having the identity of the national prefix, so it is naturally impossible to change jobs.

But he let some of his disciples join Longhan Technology.

After all, not everyone can be a scientific ascetic like him.

The supercomputer R&D department under Longhan Technology was also established.

……

Longhan Microelectronics also announced the Cailong C1 chip, but did not announce the listing plan.

But the emergence of the Cailong chip.

It's still not okay to make NV and Al nervous.

Originally, the graphics card market was a two-way competition.

If Longhan Microelectronics inserts one more foot, it will become a three-legged stand.

Then the profits of the entire market will be further squeezed.

In fact, Su Han doesn't want to come out with his own graphics card, because the comprehensive performance of Cailong C1 is not as good as NV company and AT company.

It also takes enough time to iterate and settle.

However, the profit of this area of the graphics card.

Sooner or later, he will have to intervene.

Because the price of the graphics card chip is about the same as the price of the CPU.

In other words, most of the profits of a graphics card are taken by GPU manufacturers.

That's why NV's profits are always so high.

……

With supercomputing.

Su Han was busy writing the code for the AI program.

Although Su Han's physical condition is very perverted now, tens of millions of lines of code are all handled by him alone, and it is not easy.

What's more, this artificial AI program is very complex.

If we talk about this artificial AI doctoral level.

Then the vast system is the elementary school level.

In the past, Su Han could still feel the awesomeness of this procedure.

But it wasn't until he bought the list of mid-level AI skills that he knew how he was a frog in the well.

In fact, after acquiring intermediate artificial intelligence knowledge, Su Han also has the level of independently developing artificial intelligence programs, and there are a large number of artificial intelligence template programs in his knowledge system, and they are all far ahead of this era.

But the problem is that the artificial intelligence program given to him by the system, and the program he mastered, are not at the same level at all.

The artificial intelligence program given by the system is very complex and sophisticated.

Many places are exquisite, so that the current Su Han is also amazing.

In fact, industry knowledge is like a Lego toy.

Anyway, the accessories have everything, and what can be spelled out in the end depends entirely on personal brains.

People with big brains can combine a lot of incredible things with Lego toys.

And some people can't put together anything but a wall.

But Su Han knew that if this artificial AI program could really be completed.

Then in the computer industry it will be no less than a nuclear explosion.

This AI program is invaluable.

It was impossible for Su Han to leave this program to his subordinates.

So he had to be a little tired to write it himself.

What he regrets most now is that he didn't start earlier, it will take at least a year to write, and it will take time to debug, and he doesn't even think about it for more than a year.

In order to increase the speed of writing.

Su Han prepared two computers.

Two keyboards.

Open the bow left and right.

The speed of code generation is no different from that of screen swiping.

But even so perverted.

Su Han's average is about 100,000 lines per day.

The main thing is that these codes are much more difficult than vast systems.

Some of this logic can only be understood if Su Han's current super brain is thought about.

Secondly, Su Han is busy now, and he can't be a shopkeeper.

I'm afraid that this work will take longer to complete.

……

July.

The first phase of Longhan Automobile was officially completed.

Today, Longhan Automobile can build more than 400 cars per day on average.

The production of 100,000 units will be completed within the year.

It shouldn't be a big problem.

The key is that the second phase of the project is also in full swing.

During the construction of the second phase.