After the layoffs, I became a tech godfather

Chapter 171 Chip and instruction system design completed

"Do you need expert assistance in the field of chip design?" asked Hao Cheng.

Hao Cheng has learned a lot about chip design during this period of time.

However, if one were to judge by how strong his abilities were, then that would certainly not be the case, even if his thinking ability and various abilities were further enhanced.

First, his mind was not fully devoted to this aspect, but was mainly on fuzzy and general mathematics.

Secondly, the study time is indeed still short, less than a month.

Third, even in the chip field, his main energy is devoted to micro-consciousness chips.

During the two days when Xiaosha was upgraded, Hao Cheng had been pondering the chip design that combined the micro-consciousness AI after realizing that he could not observe anything. He had also turned over the adder that had been upgraded before countless times.

Every time he read it, he would gain something new, and his related abilities would improve step by step. The design of simple micro-consciousness chips was no longer a problem for him.

But this is useless because today's chip manufacturing standards, whether it is photolithography of silicon semiconductors or chemical vapor deposition, arc discharge/laser ablation of carbon nanotransistors, cannot meet the manufacturing needs of micro-consciousness chips.

As Xiaosha said: The micro-consciousness chips currently available must still be based on binary input, but the two-dimensional concept of existing chips is transformed into a three-dimensional concept.

And this requires the current chip manufacturing foundation.

That's why he asked a question.

[Not at all. If you want to manufacture high-end chips, you do need relevant experts. But now I am just trying it out to demonstrate my ability. So, the chip design knowledge I have is enough.]

The most obvious feeling Hao Cheng had was that Xiaosha's reply had more modal particles and degree words. If it were in the past, Xiaosha's reply would definitely be "not necessary" without "completely" or "sufficient".

Hao Cheng nodded and thought: It really looks more like a human.

But this is not important. Hao Cheng wants to see to what extent it can assist him in designing chips.

The biggest difference between micro-consciousness chips and existing chips is actually the difference in dimension. Current chips, even with a stacked architecture, are two-dimensional, and the purpose of stacking is only to increase the area.

For example, if the area of one layer is 1, when stacked together, it becomes 2, but the size of the chip is still 1.

However, the concept of the micro-consciousness chip is first of all a network, solid, and three-dimensional.

[There are many ways to solve this problem. Based on the current technical level, [carbon nanotube silicon via] technology can be used to solve the I/O limitation problem between layers in three-dimensional integrated circuits...]

Xiaosha spoke first and threw a pile of papers and reference materials at Hao Cheng.

Hao Cheng read it eagerly.

"However, this technology can only solve the input and output problem. The core of micro-consciousness is 'the judgment and cohesion of data and intentions'. Even if it is created, it is just a copycat, with only the form but no substance."

Judging from the current paper, Hao Cheng's thinking ability is extremely fast. He understood the relevant things in a short time. At the same time, he raised this question.

[This is just hardware. Hardware is of course the form. As for the connotation, you need to design a mapping system, that is, the instruction system. The system is the soul! ]

"What do you think about the design of the instruction system?" Hao Cheng asked directly.

[Please, I'm just an AI! ] Xiaosha said:
[My design ability depends on the upper limit of my knowledge base. I cannot design anything beyond the capabilities of my knowledge base. This chip is obviously brand new, and my knowledge base does not have the corresponding capabilities.

[However, if you have any ideas, I can assist you in doing the basic work or expanding the work. ]

Xiaosha also explained that the reason why it can do the expansion work is because the ability of [abstract process] has been improved, so it has the ability to "create something out of nothing" and can "generate" things that do not exist in the knowledge base or have never been mentioned.

However, there is no guarantee whether the "produced" thing is correct or usable, and it needs to be verified.

Hao Cheng roughly understood the logic behind it. It was like a person who suddenly had an idea. Whether the idea was correct or feasible needed to be verified through practice.

"Well, the current instruction set systems are all fixed instruction sets, such as the Add instruction to implement arithmetic addition, lsl and asl to implement left shift, etc., but ours is obviously not so fixed." [Yes, so I wonder if we can design the instruction system to be similar to the "abstract process". ]

Xiaosha began to make random connections, forcibly linking these two originally unrelated concepts together. If it were traditional AI, it would definitely combine them based on probability, but Xiaosha, this is a thinking process.

"The abstract process cannot be designed." Hao Cheng denied it.

……

[What about the mapping of resource set → capability set? ]

……

Hao Cheng finally found a way to use Xiaosha at this stage, although sometimes the suggestions he made were very nonsensical and totally unfeasible.

However, its ability to spread is unparalleled. In a short while, it spread to more than a dozen places.

When people have no ideas, no matter how hard they think, it is difficult for new things to come out of their minds. But Xiaosha is different. It comes up with new things one after another.

Regardless of whether it is feasible or not, it is really great to inspire inspiration!

It didn't take long for Hao Cheng to find a solution that seemed very feasible.

That is, [resources] are regarded as a set, [capabilities] are regarded as a set, and the process mapping in the middle is regarded as an instruction set system.

……

[Master, the prototype of the instruction set system has been completed! ]

"Finished?" Hao Cheng just grasped a little idea in his mind and was about to think about it when Xiao Sha reminded him. He also hurriedly asked: "Have you replenished the things?"

[No, these are all your thoughts, Master. I just organized them! ]

Xiaosha showed Hao Cheng the core principles of the instruction set system, the upgrade training methods, etc.

Hao Cheng was also stunned for a moment. These things were indeed his thoughts just now, and he threw them to Xiaosha all at once.

However, when throwing, there was no pattern at all, and in the end, it was like a tangled ball of thread that could not be untied, a mess.

Just like just now, Hao Cheng hadn't sorted out all these things in his mind.

"Have you sorted it out?" Hao Cheng confirmed again.

[I see, isn’t this simple? I’m good at this!]

Hao Cheng understood instantly and discovered another extraordinary ability of Xiaosha.

"Well, let's call it pulling the thread!"

Hao Cheng nodded. This ability is very useful: human thinking is not inertial. Sometimes it diverges in an exaggerated way. Sometimes, a brilliant idea comes to mind suddenly.

But this kind of thinking is either fleeting or forgotten in an instant and cannot be recalled.

Or it would be like Hao Cheng today, seemingly having a lot of ideas, but it’s a mess.

And now, leave it to Xiaosha, it can help you sort it out.

……

In this way, a complete set of chip logic, from the basic chip design to the instruction system, was born with the assistance of Xiaosha.

(End of this chapter)