novelones Chapter 245 Breakthrough Strategies for Virtual Machines
Chapter 245 Breakthrough Strategies for Virtual Machines
A near-suffocating silence fell over the scene.
The bald, middle-aged man's words were like a thin knife, slicing Xiao Zhou and his colleagues, who were just struggling to hold on, in two.
Su Chen was not just facing Synopsys, but the network behind Synopsys that had been woven for more than 20 years – four top contract manufacturers, tens of thousands of fab suppliers, the Boston Global IP Council, and thousands of industrial simulation application developers.
Another name for this network is **"Ecology"**.
Outside the small bar window lay Shanghai on a summer night, with the lights of Hongqiao stretching out in patches until the sky and earth became indistinguishable. Inside, however, it was so quiet that the faint hiss of the air conditioner vents could be heard.
Just when the silence was almost overwhelming, the 45-year-old algorithm engineer sitting in the corner slowly put down his small wine glass.
He raised his eyes, his voice not loud, but clearly reaching everyone's ears.
"Actually—there is a way to bypass Boston."
Several pairs of eyes turned to them simultaneously, filled with astonishment.
The bald man raised an eyebrow, not immediately refuting, but instead picked up his glass and took a sip, as if to see what this senior had to say.
The veteran algorithm engineer wasn't in a hurry. He slowly pushed the cup to the center of the table and then uttered the word.
"Signature-compatible container".
Xiao Zhou was stunned for a moment.
"container?"
"That's right." The senior algorithm engineer nodded. "Don't forget, there's something called a 'virtual machine' in this world. Boston doesn't allow global foundries to approve the simulation output of Qiming Layer 2, and it doesn't allow our algorithm output to be understood—so Weilan can do the opposite and create its own 'container' within the Qiming Layer 2 OS."
He paused, then spoke very slowly.
"This container is specifically designed to recognize the syntax of the Sentaurus node library, specifically to receive requests from the contract manufacturer's approval process, complete the translation within the container, and then send the data that the contract manufacturer can understand back from the other end of the container."
"So—" Xiao Zhou's voice trembled, "what the contract manufacturer is seeing is still the familiar Sentaurus approval data stream?"
"right."
"But what's actually running underneath is the Qiming super engine?"
"right."
Silence fell once again.
This silence was no longer a silence of despair, but a near-calming silence of "Is this really possible?"
Xiao Zhou subconsciously looked at the bald man.
The bald man slowly put down his glass, looked at the old algorithm engineer, and the sarcasm on his face faded a little, but only a little.
"Old Zhang," he called the other man by name, "theoretically it's possible. But do you know how much work it would take to actually build this 'container'?"
Old Zhang remained silent, waiting for him to continue.
「Sentaurus节点库的语法,从1999年那一版到今天,翻新了多少次?新思在它上面叠了多少层私有扩展?」秃头大叔伸出手指,一根一根地数,「2003年的高阶多物理场耦合扩展包、2007年的统计变差扩展、2012年的非线性自洽叠代扩展、2016年的机器学习参数补偿……一个一个全部要翻译。」
"And that's just Synaptics," the bald man continued. "Cetronics, with its waveform signature extensions, power integrity extensions, and thermal noise extensions, adds up to several hundred interfaces. Ansys and Siemens each have several hundred more."
"If you want to create this 'container,' it's like retranslating the small mountain that the four top EDA companies have built up over the past twenty years."
Xiao Zhou's face paled even more.
"Moreover," the bald man dropped the bombshell, "if we rely solely on software to do this, how inefficient will it be? So inefficient that the contract manufacturer can't afford to wait. The contract manufacturer's approval process is charged by the hour. If your container is half or twice as slow in the middle, the contract manufacturer will kick you off the list immediately."
Xiao Zhou nodded subconsciously. He understood this.
"To increase efficiency, you need supporting hardware," the bald man concluded. "Software acts as the translator, and hardware provides the pathway. In other words, Vilan not only needs to write this container in the OS, but also needs to reserve a dedicated 'signature data pathway' for it in the chip. Only with the cooperation of software and hardware can this container run."
"The money and effort involved," the bald man finally shook his head, "I don't think Vivian can afford it."
Old Zhang listened quietly, a faint smile on his face the whole time.
He waited until the bald man finished refuting everything before slowly picking up his wine glass, looking at the lights outside the window, and softly uttering a sentence.
"Old Li," he called the bald man's name, "don't forget what happened to those companies that went against Wei Lan."
The bald man didn't answer immediately.
"In 2024," Old Zhang continued, "Magwel was relentlessly attacked by NewThink for twenty-two years. Who finally caught it?"
"In 2025, who will place Mr. Shen Shouhong's letter from 1995 on your table?"
"On June 15, 2026, Qiming's first floor tracked that curve on the Jinshajiang F7. Who made four of the world's top EDA firms simultaneously uneasy?"
He took a small sip of wine with each question he asked.
"Old Li, I actually think Su Chen can do it." Old Zhang finally put down his wine glass. "Because he has already accomplished so many things that were originally impossible."
Xiao Zhou nodded slightly beside him.
A thought popped into his mind: "Old ginger is spicier."
This unassuming 45-year-old algorithm engineer, in just a few words, first presented a breakthrough idea that stunned everyone, and then used three real events that happened in the past year to gradually dispel the feeling of "this is impossible".
The atmosphere at their table gradually shifted from "Lin Xuan—Su Chen—are doomed this time" to "Maybe there really is a way out."
But the people at this table were unaware at that moment—
At the same moment they were drinking in their tavern, the sixth-floor east unit of the main building of the Qifeng Basic Algorithm Research Institute in Zhangjiang South, Pudong, Shanghai, and the Weilan R&D Center in Building 3 of Sunqiao, were brightly lit.
The hundreds of engineers responsible for developing the Qiming Layer 2 OS were sitting in their seats, letting out a long sigh of relief.
One of them, an engineer sitting in the third row by the window, turned to his colleague and whispered:
"Thank goodness."
"Fortunately, the boss didn't let Qiming's second floor go directly down the 'pure Sentaurus approval stack' route back then."
"Otherwise, Boston's statement today, 'We will not accept the simulation output of Qiming's second floor,' would mean the hundreds of millions we burned through in the past year—"
He made a gesture and dropped down from the air.
"It was all for naught."
The colleague next to him nodded.
This is a fact.
In this year, Vilan did not follow the old path of Synaptics and Cadence, which relied solely on the Sentaurus node library to connect with global contract manufacturers' approval processes.
Wei Lan is taking the following path—
"Big Heart" dual-engine architecture.
The name was personally chosen by Su Chen at an internal meeting in December 2025.
The so-called "big heart" refers to the Qiming Big Engine—based on Vilan's independently developed third-order nonlinear extension algorithm, formal verification algorithm engine, 180 core patents, and a pure Chinese algorithm stack, running on a large hardware core.
The so-called "careful" refers to the Sentaurus small engine—based on the 1999 Sentaurus compatibility license that was just acquired for $7 million on July 12, and Magwel's three MEMS simulation patents, running on a small hardware core.
Xiao He only acts as an intermediary.
It doesn't actually participate in the calculation.
It is only responsible for receiving requests from the contract manufacturer's approval process, translating the Sentaurus syntax into an internal language that Qiming Engine can understand, and then translating the results calculated by Qiming Engine back into the Sentaurus syntax and sending them back to the contract manufacturer.
The beauty of this architecture lies in—
It can be compatible with the Sentaurus approval process, and gradually transfer the "OEM-EDA" network that NewThink has accumulated over the past 20 years to Vilan's own body.
Alternatively, one could completely shut down the Sentaurus engine and only run the Qiming engine, creating a purely Chinese-based signature stack based on the Qiming Layer 2 interface specification, and then operate independently.
Moreover, thanks to the Sentaurus mini-engine, Qiming's two-layer OS can directly understand and run industrial simulation tasks that could originally only run within the Sentaurus ecosystem through this mini-core pathway.
Su Chen's "Big Heart, Small Heart" strategy put the second floor of Qiming in a position where it could advance or retreat at will.
"What's even more amazing," the engineer in the third row continued in a low voice, "is that we hold the exclusive global rights to the Sentaurus compatibility license from 1999, as well as the rights to use Magwel's three MEMS simulation patents in mainland China."
"In other words, as long as we avoid Synaptics' proprietary extensions after 2003—such as the High-Order Multiphysics Coupling Package, the Statistical Variation Package, the Machine Learning Parameter Compensation Package, the Kintex Waveform Signature Extension, and the Ansys Turbulence Coupling Package…"
"As long as we don't touch them," he said, emphasizing each word, "it's perfectly legal for Vivian to use this 'big heart' trick."
The colleague next to him nodded.
"Even if Synaptics jumps up and Boston Consulting Group jumps up, they can't sue us. Because we're using our own legal licenses, and we haven't infringed on any of their extensions."
"However—" another engineer frowned, "we don't have access to those extensions from 2003 onwards. In terms of the finesse of multiphysics coupling, our Sentaurus engine is significantly weaker than Synaptics' latest version. What should we do about that?"
"It's alright." The engineer in the third row smiled. "The real calculation isn't done by the Sentaurus engine, but by the Qiming engine. The smaller engine only acts as an intermediary; its weakness doesn't affect the calculation results. Its weakness only means a slightly slower translation speed, while the hardware IP for the approval data path has already achieved nanosecond-level processing speeds, so the slowdown is only limited."
Several engineers around him nodded in agreement.
This is another wonderful aspect of the "big heart" concept—
It compresses the Sentaurus system into its small core, turning it into a translator rather than a calculator. A slightly weaker translator is harmless. The real power resides in the Enlightenment engine.
"The boss is really brilliant," a 30-something engineer remarked softly. "When we decided on the 'Big and Small Care' policy six months ago, who among us could have imagined that Boston would actually jump out today and slam the door shut on the global approval process from the outside?"
"The boss probably foresaw this a long time ago," said the engineer in the third row. "That's why he pushed back the 'big heart, small heart' architecture six months in advance. What Boston said today would be a thunderbolt to others, but to us—it's just rain."
The group exchanged a smile.
The atmosphere in the meeting room shifted from "What should we do?" to "We're ready."
but--
In this relaxed atmosphere, an engineer in charge of intellectual property rights frowned.
"However," he began slowly, "we still face risks at the OS level."
Everyone turned to look at him.
"The 'Big Heart' doesn't infringe on any hardware rights, that's fine. But the 'signature-compatible container' written in Qiming's second-layer OS directly reads the syntax of the Sentaurus node library and directly receives the data sent from the foundry side—"
He carefully considered his words.
"This matter is in a gray area of ambiguity at the patent level."
"There's no precedent for this anywhere in the world. If Boston really wants to mess with us, they can take this to court. Even if they don't win in the end—"
"They can drag it out, and they can wear us down too," a senior engineer chimed in. "What has the Boston Global IP Committee relied on for the past twenty years? Dragging it out. Litigation is never about winning or losing, it's about who can't hold on any longer."
The atmosphere in the meeting room became more tense.
"This is something we'd better not do directly," the intellectual property engineer said, expressing his concern.
"What should we do then? If we don't write 'heart' on the container, then only half of it will be there," a young engineer exclaimed anxiously.
Just then—
A young engineer who had been sitting at the very back and hadn't spoken until now smiled and slowly began to speak.
"You've forgotten one thing."
Everyone looked at him.
"Qiming Layer 2 OS is an open-source OS."
He pronounced those words, one by one.
The meeting room fell silent.
"Boston wants to sue? Sure," the young engineer continued with a laugh. "Let them sue those anonymous 'netizens' who submit patches in the GitHub repository."
"What does Wei Lan want with us? After all—"
There was an almost mischievous glint in his eyes.
"**Qiming Layer 2 OS itself does not 'naturally' support the Sentaurus signing process.** That container was written by a 'grassroots expert' in the Chinese EDA community, who anonymously submitted it to the open-source repository."
"Weilan just happened to merge this PR department," he said with a smile and a shrug.
In the conference room—
A moment of silence.
Immediately afterwards—
"Hahahaha—"
Hundreds of engineers burst into laughter, the sound echoing from the R&D center's conference room all the way down the corridor.
They finally figured out the key point.
Yes.
Qiming Layer 2 OS is open source.
This means that anonymous, unnamed "grassroots experts" in the Chinese EDA community can spontaneously write a "Sentaurus signing compatibility patch" and submit it to the open-source repository of Qiming Layer 2 OS.
Vilan is just a merged code.
Vilan did not actively develop any "infringing" content.
The blame can be entirely shifted to the collective term "Chinese EDA community experts".
And Boston wants to sue over this collective term?
Whom can it sue?
It sued anonymous accounts on GitHub? It sued the group avatars on Chinese EDA forums? It sued the thousands of graduate students at Beijing University of Posts and Telecommunications, Tsinghua University, Fudan University, Shanghai Jiao Tong University, Huazhong University of Science and Technology, and the School of Microelectronics at the University of Electronic Science and Technology of China who work overtime every night writing open-source code?
It cannot be persuaded.
It can't be sued by anyone.
The hundreds of engineers in the conference room looked at the three characters "Big Heart" written by Su Chen himself on the whiteboard and felt a sense of respect.
This architecture was finalized six months ago.
I received the Sentaurus compatibility license two weeks ago.
Two days ago, the five committee members in Boston launched the initiative.
today--
Once a move is made, there's no going back.