Chapter 34 Let's Sneak In First
In the software industry, white box testing, simply put, is using exhaustive testing methods to run through all the code. If you run through everything, you can roughly understand all the logical structures and pathways.
Upon hearing those four words, everyone showed different reactions—except for Zhang Ke, who was completely confused.
Even Pan Xiaojun, who had some programming knowledge, didn't fully understand. He knew white box testing was part of software testing, but he didn't get why a cleanroom was needed.
Among the three UMC engineers, two initially had the same confusion, but they quickly figured it out and stared at Su Yuanshan in shock.
Only Lin Yonggu immediately understood Su Yuanshan's true intent upon hearing "white box testing."
In the chip manufacturing industry, the only use of cleanrooms for "white box testing" was for running microcode.
Reverse engineering a CPU, in principle, is quite simple: open the chip's casing, deduce the mask layout from the transistor arrangements, and reconstruct the microcode—
Microcode resides in high-speed memory, translating machine instructions into detailed sequences of electronic operations, controlling logic gates to produce 0s and 1s.
While transistor layouts can be physically seen, microcode cannot.
Thus, one engineering team describes the functions of the microcode while another independently recreates microcode with the same functions without referencing the original—this is the real white box testing performed inside cleanrooms.
Just last month, AMD had launched their AM386 this way—hence the name, 386.
Lin Yonggu looked at Su Yuanshan's absurdly youthful face, catching the faint, calm smile there, and knew he had guessed correctly.
He took a deep breath and asked, in somewhat broken Mandarin, "You want to learn from AMD?"
Su Yuanshan nodded. "That's the plan."
"It's very hard," Lin Yonggu said with a hearty laugh.
Since meeting Pan Xiaojun, he had heard nonstop about Su Yuanshan—how this eighteen-year-old had world-class talent in programming and circuit design.
But having talent didn't mean unlimited ability.
Lin Yonggu shook his head, smiling. "AMD took three years to reverse-engineer the AM386, and the 486 is even harder. You know, they were already seasoned veterans."
Su Yuanshan also smiled.
That last point deeply resonated with him—
AMD had built its reputation by reverse-engineering Intel's chips. Founder Jerry Sanders once boasted, "We want to be Intel's second supplier, but we can't wait that long!"
Meaning: "If you don't license to us, we'll copy you."
And the most infuriating part?
AMD's clones often turned out better than Intel's originals—even winning military contracts.
Calling them veterans at reverse-engineering wasn't an exaggeration.
"Still worth trying. What if it works?" Su Yuanshan said with a grin.
"But let's not talk about that now. You must be tired from traveling. Get settled first—
You weren't planning to just grab the CDs and leave, were you?"
Lin Yonggu chuckled. "Of course not."
"Good. This way, please."
...
Uncle Zhang and Pan Xiaojun escorted the three UMC engineers to the foreign guest hotel upstairs.
Su Yuanshan saw them off at the door, then returned to Qin Weimin's office.
Xi Xiaoding also abandoned his luxurious 800-yuan office chair to join them on the guest sofa, waiting for Su Yuanshan.
"So, how did it go?" Xi asked.
"I casually brought it up. Judging by their reaction..." Su Yuanshan touched his lips and peeled off a bit of dry skin, smiling slyly, "We'll sneak in if we can. If not, forget it."
Both Qin Weimin and Xi Xiaoding laughed.
When analyzing UMC's situation, Su Yuanshan, Qin Weimin, and Xi Xiaoding had spent a long time discussing and combining their insights with Su Yuanshan's memories.
The conclusion was clear:
UMC's chip design division was under heavy pressure—they desperately needed to develop a breakthrough product to secure their place.
Gordon Moore, Intel's co-founder, had proposed Moore's Law, which had proven accurate for over twenty years.
Given this forecast, UMC, whose core business was wafer fabrication, had to consider whether their investments could keep pace with Moore's Law.
Meanwhile, their rival TSMC had committed everything to building wafer fabs, aggressively catching up.
Under such pressure, UMC had no choice but to double down on its foundry operations.
In Su Yuanshan's memories, UMC had indeed begun spinning off its chip design divisions around 1995, fully committing to becoming a pure foundry.
One of the spin-offs would later become the legendary MediaTek.
It was also worth noting that in 1994, UMC had attempted to design a CPU based on the x86 architecture—
But because they hadn't secured Intel's authorization, they couldn't legally sell it.
That failure likely solidified their decision to exit chip design.
Here, "design" clearly referred to reverse engineering.
Su Yuanshan could sense that UMC was probably already quietly attempting to reverse-engineer Intel's 486.
Otherwise, why would they be so eager to demand technical support from Yuanxin?
At this stage—before multi-billion transistor, FinFET-based CPUs—the relationship between EDA engineers, chip designers, and wafer fabs was still very intimate.
In modern times, developing a CPU required these three groups working together daily.
Understanding UMC's desperation, Yuanxin's leadership had only one thought:
How to maximize the benefits of this cooperation.
Su Yuanshan's proposal—to request cleanroom access for "white box testing"—was officially about helping UMC.
Unofficially, it was about sneaking Yuanxin engineers into UMC's reverse-engineering process to harvest valuable experience.
After a laugh, Qin Weimin asked, "What if they refuse? Just send one person?"
Su Yuanshan nodded. "Exactly. Our process design is already ahead.
At worst, send one engineer to tweak tools and liaise with their equipment."
"And if they impulsively agree? Who do we send?" Xi Xiaoding frowned.
"You know, reverse-engineering experience is only valuable for chip designers and fab engineers—not so much for EDA engineers.
UMC's lines are top-tier. Would our people even handle it? We don't have enough manufacturing experts yet."
Su Yuanshan squinted out the window, lowered his voice, and whispered like a thief:
"Senior Brother Xi, who said we have to send only our own people? What if I 'borrow' some engineers from Huajing Microelectronics?"
Xi Xiaoding's eyes lit up instantly!
Qin Weimin slapped his thigh.
"Brilliant! I bet Huajing's Director Wang will beg you to take some of their people!"
"Shhh..." Su Yuanshan pressed a finger to his lips.
Opportunities to access the latest manufacturing lines were rarer than diamonds in China.
And now—an opportunity was right in front of them.
No way would Su Yuanshan miss it.
Even a quick glimpse by a Huajing engineer would be priceless at this stage.
Of course, all of this depended on whether UMC agreed.
Thank you for the support, friends. If you want to read more chapters in advance, go to my Patreon.
Read 20 Chapters In Advance: patreon.com/Albino1