1991 Starting from the core

Chapter 503 Third Generation IO Protocol

As the saying goes, one step ahead is a genius, two steps ahead is a lunatic. Although Su Yuanshan doesn't fully agree, he also knows clearly that in technology, if the idea is too advanced, it will take more effort to convince others. tongue. And if it was just a little bit more... it would be much easier.

For example, Su Yuanshan has never mentioned that streaming video will be one of the core audio-visual experiences of the Internet in the future, let alone that the Internet will form one information island after another due to the natural monopoly pursuit of giants.

Similarly, in the computer field, he only promoted the development of USB at the appropriate time, and it progressed in a spiral with the flash memory architecture.

But the one in front of me... is a bit interesting.

…

Luo Ming came to Li Mingliu's office this time because he suddenly thought of a brilliant idea while bragging about international bragging with Mei Jie and the relevant technical backbones of the Xinghai GPU team and Yuanxin CPU team.

"Xinghai's GPU computing power is getting stronger and stronger, and the existing PCI interface will soon become a shackle restricting the computing and data exchange capabilities of the graphics card. Therefore, everyone hopes to develop a new interface standard specifically for graphics card communication - it is said that Intel has I'm also very interested in this." Luo Ming looked at Su Yuanshan and said enthusiastically: "Mr. Shan, you know this."

Su Yuanshan smiled and nodded: "Didn't you say last time that you were planning to upgrade based on the PCI bus protocol? Just change the interface to directly connect the core data, right?"

In fact, this is the path that the AGP interface has taken in history, but AGP was originally proposed by Intel in 1997.

Since the launch of Yuanxin's CPU, Su Yuanshan has consciously only grasped the overall situation and direction of the CPU team, and did not involve daily work and technology.

Of course, this does not mean that he does not pay attention to the development of the CPU department.

Now the CPU department has been divided into several R\u0026D teams, including 64-bit processor architecture, instruction set team, and protocol department.

"So I was thinking... why do we need to develop a graphics card-specific protocol? Can't we develop a general-purpose protocol?" Luo Ming looked at Su Yuanshan with burning eyes: "If we want to do it, we can do it in a big way! For the existing I/O of the computer The O-bus is undergoing a comprehensive upgrade.”

As he spoke, he ignored the shock in the eyes of Su Yuanshan and Li Mingliu and handed the document in his hand to Su Yuanshan.

Li Mingliu sat down next to Su Yuanshan with a wry smile, watching what the hell this guy had done.

Su Yuanshan woke up first, then composed himself and started reading the documents.

In the document, even though Luo Ming wrote a lot, it is better to say that it is just an idea and idea...

In his vision, his new generation - or third generation I/O bus standard theoretically has the following characteristics.

First, its performance is faster than the existing PCI-based I/O bus.

Secondly, it will have a variety of connection methods, and you can even extend the cable to connect the interface to the outside of the chassis to achieve high-speed communication with peripherals.

Third, it is a comprehensive multi-functional bus protocol and uses a topological structure and serial interconnection. Compared with the shared bandwidth of PCI, it can provide more protocol bandwidth for multiple devices.

Finally, it will be open and cross-platform, not just storage devices and graphics cards, but even CPUs can be connected through this bus if you want.

…

Li Mingliu read it first.

"No, Lao Luo, in order to build a solid-state drive, you have to mess with the existing board bus layout? What's the difference between buying a car and rebuilding all the roads in the country?" Li Mingliu said dumbfounded: "Don't say we don't have this influence. Even if we do... we won't be able to push it in the market, right?"

At this time, Su Yuanshan finally raised his head and said with a smile: "Even Intel can't push it."

Seeing that the two bosses, Su Yuanshan and Li Mingliu, were unanimously opposed, Luo Ming couldn't help but feel a little disappointed. Although he got Li Mingliu's promise to recommend him to be promoted to P14, Luo Ming still knew that the reason why his technical level had not been improved. The main reason is that Yuanxin attaches great importance to innovation in its technology hierarchy.

He admits this is his weakness. So this time he felt that...if he could start with the interface of solid-state drives, Yuanxin could seize the opportunity to fully upgrade the next-generation interface...at the very least, it would be a kind of proof for himself.

Su Yuanshan looked at Luo Ming's change from high interest to loss, and couldn't help but laugh: "Although what Lao Li said is a bit damaging...but you have to admit that your fight is indeed a bit too big."

"However, your idea is very good, and I support it."

Luo Ming was slightly startled: "How can I support you?"

"Hmm... let me think about it."

Su Yuanshan pondered.

Frankly speaking, the reason why he looked slower than Li Mingliu was because he was really amazed by Lao Luo.

Looking at this one by one, isn't this the third generation I/O protocol proposed by Intel?

The third generation I/O protocol is 3GIO, which is rarely known to people outside the industry.

But if we talk about the upgraded version of 3GIO, then I am afraid that everyone with a little bit of computer knowledge will know it.

PCI-E was born based on the concept of 3GIO.

Although PCI-E comes with PCI, it has nothing to do with PCI. On the contrary, the AGP interface is an upgraded version of the PCI2.1 bus - and it is precisely because of the inherent shortcomings of the PCI bus that AGP was eventually replaced by the faster and better PCI-E interface.

Now that Luo Ming can actually think of this, I can only say... that he has really worked hard on the computer hardware protocol.

"In this case, Senior Brother Luo... I only look at the difficulty of the bus logic and protocol engineering. My preliminary estimate is that it would be impossible to straighten this thing out without a team of 100 people working on it in one or two years. And as you know, proposing a bus protocol Although standards will help us influence computer hardware, especially motherboards and peripheral products, if our influence is not enough, it is entirely possible that we will not be able to push forward... resulting in wasted efforts."

When Su Yuanshan said this, Luo Ming sounded a little more agreeable. He also nodded and said: "Yes, I mainly consider that the future development of computers really needs to simplify the complexity - and the current I/O bus is too We are lagging behind. If we can put it forward first, it should bring positive meaning."

"In that case, we will need to unite with other companies again." At this time, Li Mingliu interjected: "Lao Luo, it's not like you don't know that it was so troublesome for us to push a USB bus... Mr. Shan even went to Silicon Valley to visit everywhere Have you forgotten? And it's been two years now, and the application of USB is still slow... You know, it would be extremely difficult to launch a standard that eliminates all the backward I/O interfaces now!"

"Well, yes." Su Yuanshan smiled and nodded: "However, I agree with Senior Brother Luo's vision. Unified I/O standards must be the general direction in the future. However, we can only go slowly now."

"For example, we can start with the ATA interface."