Chapter 206 He Who Controls Flash Memory Controls the World
When talking about flash memory, one must first mention "memory"—the memory used in RAM modules, the running memory of devices.
In specs like "8+128G" shown on mobile phones, the first number (8) refers to RAM, and the second number (128) refers to flash storage.
In computer specs, 8G or 16G refers to RAM, while a 256G solid-state drive uses flash memory for storage.
It's well known that RAM is much faster than flash memory. This is because RAM doesn't need to retain data after a power loss, so it doesn't require a structure to trap electrical charges. Electrons can quickly pass through the gate or drain in one swift charge-discharge cycle.
Flash memory, however, is slower because it includes an extra structural layer to trap electrons, enabling it to retain data even when powered off. The state of the trapped electrons can then be detected and interpreted as a 0 or 1 by external amplifiers.
Within the architecture of flash memory, there are two main types: NOR flash and NAND flash.
NOR flash has extremely fast read speeds but painfully slow erase speeds. In contrast, while NAND flash reads more slowly, its erase speed is extraordinarily fast.
Thus, NOR flash is typically used for firmware that requires frequent and efficient reading but little erasing.
NAND flash is better suited for environments that require frequent reading and deleting of data.
Later on, when people talked about SLC, MLC, and TLC "flash memory" types, these were just variations of NAND architecture.
Just before the Spring Festival, Su Yuanshan personally went down to follow the developments of the USB standardization body—USBIF (USB Implementers Forum), which had now been officially established. Yuanxin's main participants were CPU team member Xie Xiaochun and some friends from Meijie. Since the organization's core members were based in Silicon Valley, the main drivers were Xinghai, Intel, and Microsoft.
The USB standard was already roughly finalized and was expected to be formally released around mid-year.
As with any new standard, it would take a long time from its release to full market adoption. In fact, it wasn't until USB-based flash drives appeared that the USB standard began spreading rapidly.
Precisely because of this, Su Yuanshan had asked Li Mingliu's team to establish a dedicated NAND flash research group once their work on secondary cache reached a stage of stability.
Because it wasn't just about USB drives—MP3 players, solid-state drives (SSDs), external phone storage—all of them relied on flash memory.
In the future, it wouldn't be those who controlled RAM who ruled the world.
It would be those who controlled flash memory.
...
Seeing that Li Mingliu still seemed hesitant, Su Yuanshan thought carefully and spoke very seriously.
"Right now, what limits chip technology is the fabrication process. But theory must always move ahead of the process—that's an iron rule in the semiconductor industry."
"We can't wait until we need an axe to start mining iron and forging it. We must prepare the axe in advance—even if it's blunt, even if it's crude, it's better than scrambling last minute."
He glanced around at the tech experts sitting nearby and added slowly, "How did Yuanxin make it this far? By always being one step ahead."
Li Mingliu immediately nodded. "Got it. I'll go back and organize the flash memory team right away. And if I have spare time, I'll personally oversee it too. The principles behind flash memory are simple—the hard part is the manufacturing process. But since this is still the theoretical phase, we can set aside manufacturing concerns for now. We're not trying to mass-produce it immediately, right?"
"This technology isn't widely applied yet, right?" Qin Weimin asked, glancing at Su Yuanshan.
"According to some insider info, it will be soon," Su Yuanshan said with a smile.
Indeed, the first CompactFlash (CF) cards using flash memory would be officially launched this year.
After wrapping up these three main areas of focus, only the less urgent and lower-priority technologies remained.
After all, Yuanxin had its own "incubation" system.
For instance, shortly after Linux first appeared, Yuanxin had already set up a small group to modify Linux to support the YX01 CPU. (See Chapter 96.)
That group hadn't disbanded and was later split off from the EDA department. Since then, Yuanxin had quietly kept funding them—and they were now responsible for developing the Chinese-language operating system for the YX1201 mobile phone.
Today, that small group had grown into a team of thirty to forty people passionate about operating systems.
Their leader, Mao Yonggang, often shouted the slogan: "We will become the Microsoft of Yuanxin!"
Su Yuanshan always treated anyone who aspired to be the next Bill Gates with caution.
Thus, even though this team would eventually be as important as the mobile phone division in Yuanxin's future plans, he still wasn't ready to let them operate independently—or to elevate their status too much.
Moreover, Su Yuanshan had firmly intervened to stop them from developing a full-fledged desktop system. Instead, he directed them to quietly optimize Linux for enterprise-level use, making it more suitable for development environments while improving support for the YX architecture CPU.
In short, the Linux+YX architecture for desktop systems would move toward the enterprise ecosystem.
As for home entertainment ecosystems... forget it.
—Going head-to-head with Apple and Microsoft simultaneously? Su Yuanshan didn't have that level of confidence yet. After all, building an application ecosystem wasn't something you could achieve in a few days. If Yuanxin dared to develop a Unix/Linux-based system aiming to compete in the desktop market, Apple and Microsoft would crush them immediately.
(Handheld smart mobile device operating systems, however, were another story.)
...
Finally, Chen Jing announced last year's research funding usage report.
"Last year, purely for research purposes—including salaries, equipment purchases, and experiment costs—Yuanxin spent a total of 870 million yuan. This doesn't include the establishment of the Tech Park in Shanghai or the Special Zone Internet Center," she said.
Chen Jing glanced at Su Yuanshan with a faint, amused smile.
Su Yuanshan felt a little awkward.
Originally, the pure research budget was only 500 million yuan. But after some unexpected "buy-one-get-one-free" deals from the government—like the servo motor project—plus adding a precision instruments collaboration unit, setting up a chemical lab, acquiring Lin Benjian's team, founding a Silicon Valley R&D center, and expanding various labs... the total spending had shot up by another 370 million yuan.
"According to the original plan, this year's budget should be 1 billion yuan," Chen Jing said, remembering the massive application submitted by the lithography center before the New Year and feeling a headache coming on. "But now it looks like that's not enough. This year, we still have to establish another R&D center up north."
"Let's temporarily set it at 2 billion yuan," Su Yuanshan gritted his teeth and said. "I'll be careful with spending. But I have a feeling—just on test wafers alone this year, we're going to burn several hundred million yuan."
Chen Jing was stunned for a second. Forgetting she was in a formal meeting, she lightly slapped the table and glared at him. "Pah! Jinx!"
Then—Qin Weimin, Tian Yaoming, Li Mingliu, Gao Xiaodi, and all the others who might be doing chip tape-outs this year immediately chimed in.
"Yeah, yeah!"
"Exactly!"
"During the New Year period… what kind of thing are you saying!"
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