Cairo: The Mobile Application Processor
The Mobile Application Processor Cairo TF Team. It was the name of the task force formed by Dr. Lisa Su and fewer than ten researchers to design a low-power processor. Lisa Su had proposed the name Cairo. Among the four great civilizations, it was the capital of the Egyptian civilization, which boasted one of the oldest histories alongside Mesopotamia. Lisa Su had chosen the name Cairo in the hope that the processor she and her colleagues were developing would endure as long as the Egyptian civilization itself. That was why the codename of the low-power processor the Cairo team was designing for the first time had also been set as Cairo. Yoo Jae-won had liked the name from its sound to its meaning, so he had immediately accepted Lisa Su's suggestion and made it official.
The Cairo team's laboratory was located on the outskirts of Silicon Valley, so reaching it required a ten-minute drive. At first, the Cairo team's lab had been inside the same building as ID Technology's offices. However, because the trial production of the completed processor design required the use of toxic substances such as hydrofluoric acid, the risk of leakage was too great and the space too cramped, so the lab had been relocated.
"Chairman! Welcome! We've been waiting for you." Dr. Lisa Su was as energetic as ever today. Upon receiving word that Yoo Jae-won was arriving, she had come out to the entrance of the lab and greeted him warmly as he stepped out of the car, her presence radiating confidence.
"It's been a while. I'm glad to see you're looking well."
"We were chatting just yesterday. What do you mean?"
Lisa Su's sensibilities were truly unique. She was the complete opposite of Remington, who insisted that nothing compared to meeting in person even while using ID Talk constantly. For her, ID Talk alone was more than sufficient.
"Well then, let's go inside. I'll show you that your choice was the right one!"
With an air of absolute confidence, Dr. Lisa Su led Yoo Jae-won and his entourage. Drawn in by her charisma, Yoo Jae-won and the others followed without a word of protest. The lab tour that followed was brief, ending in under ten minutes. While typical semiconductor factories boasted enormous scale, this facility was focused entirely on processor design. Only the equipment for trial runs on one or two wafers stood out; the rest was not much different from Android's Alpha Team offices. The one distinguishing feature was that all the equipment was brand new, and every time they entered a new room, the fresh scent of new machinery hung thick in the air. It was so pervasive that Yoo Jae-won found himself worrying about the health of Lisa Su and the other researchers, concerned it might be filled with environmental hormones.
"The smell of new products is making me a little dizzy. We should ventilate more often."
"Ah, since this is your first time here, Chairman, you notice it right away! We've grown so used to it that we didn't realize it was still this strong."
From her reply, it didn't seem like she had been particularly concerned about it. Once she set her mind on something, she charged forward without hesitation, and right now she was completely immersed in processor development.
"I'd like to show you the special lab with the production equipment as well, but there are security procedures and quite a few things you'd have to wear, so I wouldn't recommend it."
"I understand. With this many people entering, it would compromise the cleanroom conditions anyway. Since we can see everything clearly from here, an explanation will suffice."
"Thanks to your full support, Chairman, we were able to install ASML's latest equipment in this small laboratory."
At Yoo Jae-won's words, Dr. Lisa Su began explaining the production equipment. It was highly specialized machinery that one company essentially monopolized, and the price was enough to make anyone's jaw drop. Yet Yoo Jae-won had purchased it the moment the request came in. One notable detail was that the installed equipment was the same model used at Mirae Electronics. It was far too valuable for a place that only processed about ten wafers a day, but for Yoo Jae-won, who was painting on a grand canvas, it was no issue.
"That concludes the explanation of the production equipment. Now I'll begin the presentation while showing you the actual products."
Dr. Lisa Su led Yoo Jae-won and his group with determined strides toward the largest conference room in the Cairo team's facility. A short while later—
"This is the wafer and the fully packaged chip for the codename Cairo."
Dr. Lisa Su brought forward a transparent case containing a 15-centimeter-diameter wafer and placed it before Yoo Jae-won. It looked exactly like the wafers seen in internet articles or television news, except for its smaller size. When turned under the bright lights, it sparkled with the same brilliant rainbow hues. The packaged processor also resembled Intel or AMD products, though it was slightly smaller.
"What is the area of each chip?"
"6mm by 12mm. That comes to 72 square millimeters. The production process is 300nm."
Lisa Su's answer was crisp and precise. However, Yoo Jae-won still couldn't quite grasp the significance. The production process was at a level so rudimentary compared to the semiconductors from his previous life that it was hardly comparable. Of course, even Intel and AMD's latest products looked the same to him, but since the Cairo team was focused on low power consumption, it was difficult to predict performance just from the process node or die size.
"I'll show you the slides."
At Yoo Jae-won's request, Lisa Su signaled another researcher who had been waiting. The conference room lights dimmed slightly, and a screen lit up via projector. It displayed a chart comparing specifications across various technical categories.
"This is a comparison chart between the codename Cairo and Intel's latest model, the Pentium II 700MHz."
The chart featured Intel's most expensive Pentium II 700MHz model. Nicknamed "the blast furnace" by netizens due to its enormous heat output, it was a monstrous chip that couldn't be cooled by the stock cooler and required a high-end aftermarket air cooler. The iWorks had also been equipped with this model, and even then it had needed a water-cooling system to manage the heat. Ironically, the bold decision to install water cooling along with green LED lighting had driven iWorks sales beyond expectations, but that only underscored how difficult it was to run the Pentium II 700. Its performance, however, was undeniable. Being over 30% faster than AMD's top-tier model, it had become common sense to recommend Intel if one sought nothing but raw performance.
The codename Cairo was half a step ahead of Intel's process. While Intel's was 350nm, Cairo's 300nm was noticeably finer. The die size was about 40 square millimeters smaller than the Pentium II. Power requirements were also reduced: the Pentium II needed 2.8V, while Cairo operated at 1.5V.
"What you're most curious about, Chairman, is computational performance, isn't it?"
"Of course! I'm very curious."
Looking at Cairo's specifications, Yoo Jae-won's expectations were rising. Sensing this, Lisa Su advanced to the next slide. A table appeared showing benchmark results for both processors running the same tests.
"Hmm?"
At a glance, Cairo's performance was roughly one-fifth that of the Pentium II 700MHz. The gap looked overwhelming. If Cairo had been an x86-compatible processor competing in the market, it wouldn't have even beaten the long-obsolete Cyrix. But Cairo was not an x86-compatible processor.
"Cairo is a RISC-based low-power processor completely redesigned from a blank slate, breaking away from the x86 architecture. Since it was optimized for low power, a fair comparison requires looking at performance per watt."
The Pentium II and Cairo used fundamentally different processor architectures. As a result, the Pentium II's numbers were measured while running the Android operating system, while Cairo's were taken on a temporary test board running a mobile operating system originally developed for the Tiffany Phone, ported to match Cairo's RISC architecture. What Lisa Su referred to as performance per watt was simply the processor's benchmark score divided by its power consumption. This yielded a clear figure of performance per watt. While this metric wasn't particularly important for desktop computers that could draw unlimited power from the wall, it was extremely critical for mobile devices with limited battery capacity.
"In terms of performance per watt, Cairo surpasses the Pentium II by more than 10%."
Following Lisa Su's words, the previously blank field on the slide filled in with an impactful animation. The Pentium II scored 81 points per watt, while Cairo scored 90. This meant that under severely constrained power conditions, Cairo delivered significantly better performance than the Pentium II.
"As expected of Dr. Lisa Su!"
Yoo Jae-won applauded enthusiastically, and Kim Dae-seok along with the ID Technology executives who had accompanied him joined in.
"This achievement belongs not only to me but also to the researchers who worked alongside me and to your boundless support, Chairman."
While enjoying the applause directed at her, Lisa Su generously shared the credit with others. Of course, the accomplishment was a team effort, and Yoo Jae-won's financial backing had been crucial, but there was no denying that Lisa Su's own role had been the largest.
"I'm delighted to have succeeded at the task you assigned me, Chairman, but the problem is that no matter how I think about it, there isn't a device worthy of using this product."
Lisa Su's presentation wasn't over yet. A question that had arisen around the time Cairo's completion as a low-power processor had become imminent still remained. Cairo's performance per watt was unmatched. The issue was that by abandoning x86 compatibility to achieve this, it had become completely incompatible with all existing software. Since Yoo Jae-won owned Android, porting the Android OS to the Cairo architecture would allow the use of existing software. However, its ambiguous performance would then become a bottleneck. While its performance per watt exceeded the Pentium II's, that didn't mean unlimited power could be supplied. Cairo's maximum stable power consumption was around 25 watts. The Pentium II 700MHz, by contrast, consumed 90 watts at stock and over 120 watts when overclocked, but it delivered corresponding performance. Installing a low-power processor in a PC would therefore be foolish.
What about laptops, where low power truly mattered? Lisa Su had naturally considered that as well. She hadn't stopped at theory; she had even built a prototype laptop.
"Even as a prototype, it achieved more than twice the runtime compared to models using Intel or AMD's low-power CPUs."
Intel and AMD did release CPUs labeled as low-power, but they hadn't changed the underlying architecture. They simply lowered clock speeds and added features to reduce power further. If a rough prototype already delivered more than double the efficiency, a properly engineered version might achieve three times or more. The problem was the x86 architecture. A dedicated operating system for the Cairo processor would be required, and applications would also need to be developed separately. Even then, because it couldn't match the high performance of x86 products, its usable range would be further limited.
"If optimized well, it might handle something like ID Office 95, but gaming would be impossible."
Yoo Jae-won agreed with Lisa Su's assessment. The reason they had created an entirely new low-power processor by abandoning x86 was not to replace existing products.
"A new processor belongs in a new product."
As he spoke, Yoo Jae-won held up the Tiffany Phone.
"The Tiffany Phone? Isn't the performance a bit excessive for that?"
"For now, yes."
Even Dr. Lisa Su, an elite among elites within the ID Group, seemed unable to grasp how the device she had created would ultimately be used. In hindsight, the answer was obvious, but attempting to solve it directly had been difficult.
"Right now the LCD is small, isn't it? But imagine that LCD screen becoming two or three times larger."
"Ah, then as the amount of information displayed increases, the required computational power would also rise. It wouldn't just handle text messages anymore; it could run various programs as well."
As expected, Lisa Su was a genius who understood ten things from a single hint.
"Exactly. Furthermore, resolution will increase, the number of displayable colors will grow dramatically, and sound quality will improve. Moreover, since a mobile phone is fundamentally a completely different system from a PC, the lack of x86 compatibility won't matter."
"That's true!"
Lisa Su nodded vigorously in agreement with Yoo Jae-won's words. Cairo's fatal weaknesses were its ambiguous computational performance and its incompatibility with the x86 architecture. Yet if it were installed in a mobile phone, both of those fatal weaknesses would effectively disappear.
"So you called me here with this in mind from the very beginning, Chairman."
Yoo Jae-won nodded at Lisa Su's words. He had taken a chance by probing, and he was immensely grateful that she had come all the way to San Francisco on her own. As expected, his bet had paid off, and the results had exceeded expectations with the Cairo processor.
"It would be a shame to simply dismiss such a product as a low-power processor. So I've been thinking of a new name. How about Mobile Application Processor?"
"Mobile Application Processor... Wow! That sounds good. I had braced myself because I already knew about your naming sense, Chairman, but it was unnecessary. This is definitely beyond expectations."
Yoo Jae-won found himself unexpectedly on the receiving end of criticism. Especially since he had always prided himself on having the best naming sense, the blow was considerable.
"Hmm? What exactly was wrong with my naming sense?"
"Isn't it a bit too straightforward? The email service site is Email.com, the online software sales site is ESD.com, the computer shaped like an egg is Egg, and the scallop-shaped laptop is Shell Book."
"Hah!"
Yoo Jae-won felt a pang at Lisa Su's barrage of facts. He had thought his naming had some flair, but hearing it laid out like this made him wonder if it was too on-the-nose. This was a secret he could never tell anyone, but Yoo Jae-won's naming sense was modeled after the trends from the peak of his previous life. Back then, product names with no apparent connection had been mainstream, but over time a backlash had made straightforward names fashionable. Having been actively involved in society through AI research at the time, he had naturally absorbed that strong impression, and it was still influencing him now. At least the name Mobile Application Processor was beyond the level of naming sense Yoo Jae-won had previously demonstrated, so it received a warm welcome from Lisa Su and the rest of the Cairo team.
"This may be presumptuous, but it seems it will take quite some time for the mobile phone you're envisioning to become mainstream, Chairman."
"It can't be helped. LCD technology needs further advancement, flash memory chips need to be mass-produced, and most importantly, mobile communication services need to become widespread. But time always flows as fast as a bullet. Now is the perfect time to prepare."
Everyone nodded at Yoo Jae-won's words. It was far better to prepare in advance than to find oneself unprepared when the launch date was right around the corner. In reality, time was flowing even faster than anyone had imagined.