1991 Starting from the core

White box testing in the software industry, in layman's terms, is to run all the code through the exhaustive method. After running, you can almost understand all the logical structures and logical paths.

After hearing these four words, everyone had different expressions except Zhang Ke who was confused. Pan Xiaojun, who has a little knowledge of programming, is puzzled. He knows that white box testing is for software testing, but he doesn't know why a clean room is used.

Two of the three UMC engineers were like Pan Xiaojun at first, but they soon guessed something and looked at Su Yuanshan in shock.

Only Lin Yonggu understood Su Yuanshan's intention after hearing those four words.

In the chip manufacturing industry, there is only one "white box test" that requires the use of a clean room, which is to run microcode.

To put it bluntly, the principle of copycat CPU is actually very simple. Open the cover, reversely push out the mask according to the layout of the transistor components, and then get the microcode - the microcode resides in high-speed memory and translates machine instructions into detailed circuit operation sequences. The control logic gate completes the output of 0 and 1.

The transistor layout is visible, but the microcode is invisible. Only one group of engineers can describe the function of the microcode, and another group develops microcode with the same function without referring to the original microcode - this is what happens in a clean room White box testing.

This is how the AM386, which AMD just launched last month, was created - listen to the name, 386.

Lin Yonggu looked at the indifferent smile on Su Yuanshan's extremely young face and knew that his understanding was correct. He took a breath and asked in non-standard Mandarin: "Do you want to learn AMD?"

Su Yuanshan nodded: "I have this plan."

"It's too difficult." Lin Yonggu smiled generously. From the moment he met Pan Xiaojun, Pan Xiaojun kept introducing Su Yuanshan - so he knew that this 18-year-old boy had a program that made people jealous. A flair for design and electronic circuit design.

But having talent does not mean having unlimited abilities.

Lin Yonggu smiled and shook his head: "AMD took three years of reverse engineering to complete AM386, and 486 is even more difficult - you have to know that they are still a bunch of veterans at this game."

Su Yuanshan also smiled, Lin Yonggu's last words deeply touched his heart. Since its inception, AMD has started by copying INTEL's various chips. Founder Sanders also proudly said, "We hope to become Intel's second supplier, but we can't wait!" The implication is, brother, you can't If you authorize it, I will copy it first as a respect.

The most annoying thing is that AMD’s copycat products are actually better than Intel’s genuine products, and have even entered the military market. To say that they are veterans of copycats is not an exaggeration at all.

"You have to give it a try. What if it works?" Su Yuanshan replied with a smile: "But let's not talk about it now. You guys are tired from traveling and traveling. Let's settle down and have a rest first - you don't think about it just because you took the CD. Let's go back?"

Lin Yonggu also smiled and said: "Of course not."

"Then please."

…

The uncle and Pan Xiaojun took the three of them to the foreign-related hotel upstairs. After Su Yuanshan arrived at the door, he returned directly to Qin Weimin's office. Xi Xiaoding also left his 800 yuan luxury chair and came here to sit with Qin Weimin to receive guests. The sofa is waiting for Su Yuanshan.

"How to say?"

"I just mentioned it casually, to see the reaction there." Su Yuanshan touched his lips, tore off a piece of dry skin, and said with a smile: "Anyway, let's rub it for a while first, and let's go in as soon as possible. Don't let us pull you down."

Listening to this guy's pornographic accent, Qin Weimin and Xi Xiaoding laughed at the same time.

Faced with UMC's request, Su Yuanshan took Qin Weimin and Xi Xiaoding to analyze for a long time. Combined with Su Yuanshan's memory, the three finally came to a conclusion - UMC's chip design department is currently experiencing a sense of crisis and urgently needs to design NB products. Establish a position.

Moore, one of the founders of Intel, proposed "Moore's Law" that the number of transistors will double in two years, which has been confirmed by the development of the semiconductor industry over the past two decades. Since there is a proven rule that can predict the development trend of the semiconductor industry in the next few years or even more than ten years, UMC, whose main business is wafer fabs, must consider whether its investment in wafer fabs can support itself. Catching up with Moore’s Law.

You know, there is a peer on the island, TSMC, which invests all its resources in wafer factories. Moreover, that peer is still ambitious, has received support from many companies, and is currently catching up with itself.

Under this circumstance, UMC had to consider increasing investment in wafer fabs - in Su Yuanshan's memory, UMC did begin to divest its chip design business in 1995 to engage in professional chip foundry. Among these spin-off departments, one company chose to focus on low-end mobile phone chip design and became extremely eye-catching in later generations - it was called MediaTek.

And it’s worth mentioning that UMC once “designed” a CPU based on the This was the final word in deciding to divest the design business in 2016.

The "design" here obviously refers to the cottage.

Su Yuanshan already has a feeling that UMC is now secretly "designing" or reverse engineering INTEL's 486 using the X86 architecture. Otherwise, how could we explain that they would come to our door so impatiently and ask Yuanxin to send a technical support team? It can only be that they are working on CPUs and are not yet proficient in it, so they need all-round support from the EDA side.

You must know that this is not yet the era of multi-level structures based on FinFET and a CPU with billions of transistors in later generations. In later generations, EDA engineers, chip designers, and wafer factory engineers have been working together for a long time and advancing progress together. This is one of the reasons why a CPU is a gathering of the most cutting-edge technology and talents of mankind.

After clarifying the urgent needs of UMC's chip design department, the entire technical management team of Yuanxin began to think about how to maximize the benefits of this cooperation. In the end, it was Su Yuanshan who made the suggestion to let UMC open the clean room and let Yuanxin carry out reverse engineering. This is what he said on the surface, but the actual intention was to let Yuanxin join UMC's chip reverse engineering in order to obtain a large number of chips. Experience.

After laughing, Qin Weimin asked: "What if they don't agree? Will we only send one person?"

Su Yuanshan nodded: "Of course, we are leading in process design now. At most, we can send one person to adjust the tools and help connect the machine."

"Then what if they agree on a whim? Which team will you send over? Do we have so many candidates?" Xi Xiaoding frowned and said: "You have to know that the best reverse experience is from chip designers and wafer factory engineers. Take it - EDA engineers are of no use. Moreover, UMC’s production line is currently the most advanced. Can our people absorb that experience? Not to mention that we don’t have the talent for the chip production line at all now.”

Su Yuanshan squinted his eyes and looked outside, lowering his voice like a thief: "Brother Xi, the people we send there may not only be our current employees! I 'poached' a few people from Huajing Microelectronics, what do you think? Can I dig it?"

Xi Xiaoding's eyes immediately lit up!

Qin Weimin slapped his thigh!

"This is a good move! I guess King Huajing will always be crying and begging you to go poach their people."

"Shh..." Su Yuanshan made a gesture of silence.

There are too few opportunities to get access to the latest production lines in China, or there are no opportunities at all! Now that there is a chance in front of him, Su Yuanshan will of course give it a try.

For Huajing's wafer fab engineers, even just taking a look at it is still a gain at this stage.

Of course, the premise is that UMC is willing to agree.

[Recommendation ticket]