The Coppermine Prototype
Technopia, 1993. Ding-dong! The doorbell ringing from the entrance had never sounded so welcome. Yoo Jae-won sprang up from his seat and dashed straight to the front door. When he flung it open, a stranger stood there holding a large box, accompanied by the security guard who always watched over the house. The stranger's cap, his clothes, and the box he carried all bore the AMD logo in clear view, so there was no need to ask where he had come from.
"Hello? Chairman Yoo Jae-won, I'm Terry Micron, CTO of AMD," the man in the AMD cap said, greeting him after confirming his identity. Jae-won had expected only a delivery, yet here stood AMD's Chief Technology Officer in person. The security guard beside him gave a slight nod. Anyone visiting Jae-won's residence had to pass the guard's identity check first, and the nod confirmed that Terry's introduction was genuine.
"Ah, I'm Yoo Jae-won. Please, come in." There was no sending the man away after receiving only the box. Jae-won held the door wide open, and Terry Micron stepped inside. Once in the living room, the two immediately opened the box. The first thing they saw was the thick Styrofoam packing material, the kind commonly used for cushioning in the United States. AMD's headquarters was only a twenty-minute drive from Jae-won's house—close enough to touch if one fell forward—yet the packaging was meticulous, as though prepared for a transcontinental shipment.
"Just in case," Terry Micron said with a hint of embarrassment as they removed the packing together. It seemed he had packed it himself. Jae-won found the man's name rather fitting. He had not expected a surname like Micron, and now the CTO of AMD bore it at a time when CPU fabrication processes were measured in microns. Of course, remaining stuck at the micron level would be disastrous; the industry needed to advance into the nanometer scale.
"Oh, so this is the first CPU made with the Coppermine process." The CPU box, roughly the size of two adult fists, caught his eye first. It looked like retail packaging, complete with attractive graphics, and the AMD K6 brand name stood out prominently. Below it was the performance rating: PR166.
"The operating speed is 120 MHz. However, we are confident it matches the performance of Intel's Pentium 166 MHz model," Terry Micron boasted with self-assurance.
"Oh, that's impressive!" Jae-won responded enthusiastically. The AMD CPUs currently installed in the data center he had built in Coyote City ran at only 50 MHz. Raising the operating speed by 2.4 times in one leap was an enormous improvement. Still, it fell far short of the 200 MHz he had hoped for. Apparently, stabilizing the Coppermine process for their own use would simply take time.
Next came the motherboard. Although AMD's strength lay in backward compatibility—new CPUs could usually run on older boards—this time the significant jump in speed required a new motherboard as well. In addition, there was future electronics HPC-certified memory and a high-capacity hard drive boasting 640 MB. The two men sat side by side and began assembling the computer. Without a case, working in a nude state, the process moved quickly. They used a spare video card, monitor, and keyboard from Jae-won's house, and in no time a high-performance PC stood ready.
They installed the Android operating system and immediately ran a benchmark. The program chosen was, of course, Doom 2, currently enjoying the greatest popularity. In the setup menu, the benchmark option at the bottom triggered a demo with spectacular graphics, and the score was calculated based on frames rendered per second.
"Wow!" When the result appeared, Jae-won could not help but exclaim. Terry Micron shrugged his shoulders proudly at the reaction. The massive performance gains from a PC whose CPU, video card, and memory all carried the HPC certification mark were clearly visible. At this level of score, even the most insensitive person could not fail to notice the improvement in computing power. Above all, the graphics in games now looked dramatically more vivid.
When the first 3D acceleration cards had appeared, they could only process 300,000 polygons per second with a single texture. The recently released HPC-certified 3D acceleration cards handled one million polygons. The catch was that they had still been paired with old Pentium CPUs. Now, replacing just the CPU with one born from the Coppermine process had transformed the benchmark scores entirely.
"Over ten thousand points!" A five-digit number appeared on the monitor. Typical Pentium-class PCs scored around three thousand points on the same benchmark; this machine had shattered the ten-thousand mark in a single bound—a more than threefold improvement. Detailed metrics showed polygon processing capability doubled to two million per second, and texture processing speed had likewise doubled. If performance of this caliber became widespread, game graphics would improve dramatically beyond current standards.
Even Doom 2, which currently boasted the best graphics, looked crude to Jae-won's eyes. It lacked dynamic lighting, so firing a weapon that produced sparks only showed the effect without illuminating the area around the character. When bullets missed and struck walls or floors, there were no flying fragments to convey impact. These were not features omitted due to technical shortcomings but rather technologies removed because the computers of the time lacked the necessary power. Now that computing performance had surged, all those features that had been developed and then tearfully cut could finally be restored.
"Ah! Don't touch that. It's hot!" Terry Micron quickly stopped Jae-won as he reached out to touch the AMD CPU seated on the board, drawn by its lovable appearance.
"Ow! Hot!" But the warning came too late. Jae-won, who had already touched it, yelped at the sharp sensation and yanked his hand back, then shook it vigorously as the heat lingered. Terry Micron looked flustered. Although it was Jae-won's own carelessness, he felt as though he had failed to warn him in time. If this ended as a mere incident it would be fine, but the real problem was that if AMD fell out of favor with Jae-won, it would deal a serious blow to their business. In the computer industry, ID Technology held absolute power as the top buyer. When a security patch had been applied last year, Intel's performance had lagged slightly behind other companies' products, and their sales had subsequently declined. The drop had been small enough to dismiss, yet for a company the size of AMD—only one-tenth of Intel—it had been a meaningful figure. That was why cooperation with ID Technology mattered. The better Android OS ran, the higher consumer purchasing power would rise.
"Quite hot, isn't it?" Contrary to Terry Micron's hopes, Jae-won had become thoroughly intrigued by the hot CPU. Visually, it did not appear particularly hot. Human eyes could not see infrared, so the lack of visible heat was misleading. Still, its shiny gold-plated surface looked irresistibly appealing.
"Yes, it is only natural that heat output rises along with core clock speed. Thanks to the Coppermine process, however, we were able to keep temperatures manageable enough to raise the operating speed," Terry Micron replied as though it were nothing special. He was right; it was common sense that higher semiconductor operating speeds produced more heat. Yet the sight prompted a question in Jae-won's mind.
"Don't you use a cooler?"
"A cooler…?" Terry Micron countered with a question of his own. It was common knowledge that higher operating speeds caused electrons to cluster more densely and resistance to increase, generating more heat. That was precisely why the Coppermine process had begun—using copper, which had lower resistance, to achieve higher speeds. Conversely, controlling heat should also allow higher speeds. And controlling heat was far simpler than changing semiconductor materials or designs: simply attach a cooler. Yet AMD's chief technology officer looked puzzled at the mention of a cooler.
"We tried attaching one, but it did not yield significant improvement." He was not unaware of coolers' existence; the coolers Terry Micron had in mind simply differed greatly from the one Jae-won envisioned. Terry's cooler was palm-sized, flat, and equipped with a small fan—the kind used on server products. The cooler Jae-won referred to was a large, robust tower-type model fitted with heat pipes.
Realizing words alone would not suffice, Jae-won stood up and headed for his study. Calling it a study was generous; it was more of a workshop, filled with far more computer parts than books. When he returned, he held a tower-style air cooler in his hands.
"What is that?"
"A cooler. I had it made." Although he claimed to have made it himself, the craftsmanship was anything but crude. That was because it had been commissioned from Lightning Bolt. A company skilled enough to build bicycle frames from carbon fiber had no trouble working with thin copper plates and heat pipes. Moreover, since it was for Jae-won's personal use, President Michael Bolton had crafted it himself. This was one of Jae-won's many efforts to squeeze every last bit of performance from computers. The unfortunate reality was that despite the money and effort invested, performance had not improved dramatically—the product had inherent limitations.
"Let's install this and run the benchmark again." Leaving the astonished Terry Micron staring at a cooler far beyond anything he had imagined, Jae-won began the installation. He applied thermal grease and securely fastened the cooler to the motherboard so that its contact surface met the CPU perfectly. Then he ran the benchmark once more.
"Ten thousand five hundred points." Simply attaching the cooler had produced a five-percent performance gain. But this was only the beginning. The true value of a massive cooler revealed itself during overclocking. Jae-won boldly began overclocking. Overclocking in this era was entirely mechanical. Unlike the twenty-first century, where one could simply press a button in the BIOS, it required manipulating jumpers on the motherboard's FSB and drawing lines on the CPU's resistors with a pencil to raise the multiplier. For Jae-won, whose computer knowledge was perfect, it was an easy task; for a beginner, it would have been unthinkable.
After finishing, Jae-won powered the computer on. Although the BIOS did not recognize the altered CPU values, the system booted normally.
"It's fast!" Terry Micron, eyes wide and focused on the monitor, reacted immediately. As the man who had led development of the new K6 CPU, he had seen the Android boot screen hundreds, even thousands of times. The time it took to reach the desktop was etched into his memory. The speed now displayed—with Jae-won's enormous cooler attached and the forbidden art of overclocking applied—was faster than any system he had ever witnessed.
The benchmark program launched immediately afterward.
"Hup!" When the Doom 2 benchmark finished after roughly two minutes and the score appeared on the monitor, CTO Terry Micron let out a gasp.
"Thirteen thousand points! That's a world record." ID Group, which led the internet, had passed that trait on to its subsidiaries. As a result, Doom 2's multiplayer features had been pushed to their absolute limits. Naturally, the benchmark scoring and internet sharing functions had been developed as well. People who wanted to show off their systems existed in every era, so tens of thousands had registered their benchmark scores. Yet this roughly assembled HPC machine, without even a case, had completely surpassed the records built by so many users. Even more astonishing was that the only change had been the addition of a single cooler—no alteration to the CPU manufacturing process. Such a simple solution had existed all along. Terry Micron suddenly felt impatient; he clearly wanted to return to his own laboratory and collect proper data.
"I'll give this to you as a gift." Jae-won removed the cooler attached to the board and handed it to Terry Micron.
"Thank you!" Terry Micron accepted the cooler with both hands. Westerners typically did not make ritualistic refusals when receiving gifts, yet Terry clutched it to his chest as though it were the One Ring and he were Gollum, determined never to let it go. Jae-won could not help but laugh at the sight. The parts AMD had delivered today alone easily exceeded three thousand dollars in value. Seeing the man so delighted over a custom part whose material cost was under a hundred dollars made him glad he had given it away. Besides, coolers were components that would naturally emerge alongside CPU advancement. If this gift exchange helped popularize high-end coolers earlier than otherwise, that would be a good outcome as well.
"Hmm, then shall we make a proper one?" After seeing Terry Micron off—still cradling the cooler protectively—Jae-won returned to his study. Now that he had gathered HPC Class CPUs, it was time to combine them into a proper new product. He picked up a pack of A4 paper and a 4B pencil with a smooth writing feel. One might wonder why a pencil when a high-performance computer sat right there, but drawing with a mouse had been something Jae-won had never succeeded at despite his efforts in his previous life. A tablet had helped somewhat, yet it still fell short of pencil work. That was why he always used a pencil for sketches and concept art.
Jae-won's hand moved without hesitation. Each stroke produced a line, and the lines gradually formed a shape—a rectangular, flat object. The corners were rounded, and inside that rounded rectangle another rectangle was drawn with a small margin. Below the rectangle, he added a fingernail-sized square button. Vertical lines gave it a three-dimensional feel.
"There." The result was an extremely simple drawing made with just a few lines. Yet for Jae-won, it was a shape that stirred deep nostalgia. Still, this was not yet a smartphone. Fitting PC-class CPUs and other components into something the size of a palm was physically impossible. It shared the smartphone's form, but at a much larger scale—a tablet PC.
"If I make the tablet PC first and then simply shrink it, that will be the smartphone." Jae-won's roadmap was straightforward: complete the tablet PC, then reduce its size to create the smartphone. The sketch he had drawn was essentially identical to a smartphone's basic shape. He had included the prominent home button at the bottom of the screen, the volume controls and various ports on the side, the 3.5 mm headphone jack at the top, and the charging USB port beside it. Then he suddenly remembered something he had forgotten.
"Huh? Come to think of it, USB doesn't exist yet." Current computer motherboards were filled with nothing but serial ports. He had examined the latest motherboard Terry Micron had brought in the box, and it too lacked USB ports.
"Ugh, does that mean I have to make this too?" USB had been created so that even computer novices could easily use peripherals—just plug it in and it worked, with nothing simpler. In contrast, the serial ports of the time required carefully matching jumpers for each device to avoid conflicts.
"If I'm going to make it, I might as well hurry!" Jae-won's thoughts raced. HPC-certified parts would soon flood the market, triggering a natural generational shift as users upgraded their computers. Supplying USB ports in advance would dramatically increase adoption rates.
"So the leisurely days are over, then?" USB ports were only the beginning. For a tablet PC without a keyboard or mouse, he would need a touchscreen interface, along with a much lighter version of the Android operating system to support it. He had never been idle, but from now on, he would barely have time to breathe.