Returning to Dominate The World With My Knowledge System
Chapter 62: Progress of Absurdities

Chapter 62: Chapter 62: Progress of Absurdities

To understand what Tyler had just accomplished with the DRAM, you had to first understand what DRAM actually was.

Dynamic Random Access Memory—DRAM—was the heart of any computing system’s active operations.

Whenever you turned on a computer, your operating system didn’t run from the storage drive. No, it was first loaded into DRAM.

That’s because DRAM offered ultra-low latency, allowing data access times in the nanoseconds range, which allows the CPU to execute processes, switch tasks, and handle real-time operations without delays.

In short, DRAM wasn’t where your files lived, it was like where your computer thinks.

By 2010, the DRAM capacities in regular consumer computers typically hovered between 2GB and 8GB. High-end workstations might push 16GB. Servers’ ran with between 32GB and 128GB of DRAM.

Anything above that was not just rare, it was industrial or government-grade and expensive.

But Tyler had achieved the inconceivable. He had created DRAM nodes of 16TB each.

Sixteen terabytes in a single stack. That was more than a hundred times the memory used by most powerful supercomputers in his time.

To make that work, he didn’t just use existing architecture, he rewrote it by combining multiple emerging technologies.

Some of the technologies were not even conceptualized by mainstream companies yet.

He used 3D stacking, advanced etching, SOI substrates, and embedded micro-controllers in every DRAM node. It was the equivalent of giving each stick of RAM its own miniature brain.

And now, he had started building something even crazier, which was the storage-tier DRAMs.

In most computers, storage and memory were separate. You had volatile memory, DRAM, for active computing, and you had non-volatile storage hardwares like hard drives and SSDs for long-term file retention.

Solid-State Drives (SSDs) were already a huge leap forward compared to hard drives. But even SSDs had limitations, which are latency, wear-out, controller bottlenecks. That’s where NVMe came in.

NVMe, short for Non-Volatile Memory Express, was a communication interface and driver designed to maximize the performance of SSDs by allowing them to interface directly with the CPU via the PCIe bus.

It reduced latency, massively improved data throughput, and replaced the decades-old SATA and AHCI standards.

But there was just one major problem: NVMe didn’t even exist yet.

In Tyler’s current time—late 2010—NVMe wasn’t a commercial product.

The protocol wouldn’t be introduced until 2011, and the first NVMe-based SSDs wouldn’t hit the market until 2013. And even then, those drives had capacities ranging from 128GB to 512GB.

The idea of building 16TB NVMe-like storage using DRAM technology would’ve sounded like lunacy.

Yet here Tyler was, doing exactly that.

He had created a non-volatile DRAM storage node with a single unit capacity of 16TB. It retained the access speeds of RAM while achieving the storage durability of NVMe.

And more impressively, he had done it by integrating capacitor-free Z-RAM structures with voltage-buffered refresh stabilization—making the data resilient even when power was lost.

In short? He had created a DRAM-based NVMe hybrid, decades ahead of schedule.

This was exactly the reason why Tyler was all smiles of satisfaction with the successful creation of the DRAMs yesterday and the second version of it today. It was nothing short of a major achievement.

From the Valkyrie-X GPUs to the memory towers and now to these hybrid storage nodes, everything he had built so far would be considered alien technology if anyone else got their hands on it.

That was the danger too.

If the wrong eyes found out what he was building and what he had already built, it wouldn’t just attract interest. It would paint a target on his back. The kind of target that led to unmarked graves and "accidents."

This was why Tyler had been so secretive with everything he had been doing so far. If it wasn’t for the NDAs that the workers had signed, he wouldn’t be telling them everything he had told them so far. Heck, they wouldn’t even be at the fab and working for him.

Of course he was aware that the document wasn’t all binding but Tyler wasn’t too worried. Yes, it’s true that they could breach the contract and sell out what they know to others, but that was extremely unlikely.

A very deep background check had been done on the applicant before they were brought to Cameroon. So, the chances of them betraying was low.

Tyler turned to his team, smiled in appreciation and congratulations, and told them to continue.

He explained to them that he needed 11 more of the non-volatile DRAM-based NVMe units—each 16TB—for a total of 192TB in storage.

This would be enough to house the AI’s training datasets, log patterns, redundancy backups, and perhaps even multiple cloned instances if required.

The team nodded without hesitation and immediately got to work. If Tyler wanted 12, they would give him 12. And they would make sure every single one passed perfection-level diagnostics.

Fueled by adrenaline and wonder, they resumed their work.

Layer by layer, chip by chip, they stacked and bonded the wafers. The first node had already been created. Now they pushed forward, armed with momentum and confidence.

But as the hours dragged on, and accuracy slowed under the strain of physical and mental fatigue, only two more nodes were completed by the time evening settled.

The total stood at three out of twelve.

Still, the team pushed on. Some refused to sit and others insisted they could go another few hours. They were addicted to the progress and determined to finish everything before the sun came up.

But Tyler raised a hand.

"No."

This caused everyone to stop what they were doing.

"I appreciate your passion. But we’re not rushing this. Not with something this delicate," Tyler said, his tone gentle but firm. "A single slip-up can fry an entire unit. You’re working on the most advanced hardware in the world. I need you rested, not reckless."

Hearing this, the team grumbled in protest, but they couldn’t argue with the logic. And deep down, they respected it.

They knew that Tyler wasn’t treating them like machines. He was actually protecting their dignity.

As for Tyler, the reason why he told them all to go home wasn’t anything special. It was just as he had said.

Though he appreciated their enthusiasm but he can’t have them working overtime. He needed to have them at their best as what they are working on are extremely delicate, complex and difficult. It would be disastrous if someone was to ruin a nearly finished DRAM due to lack of rest.

The team smiled and decided to rest for the day after not being able to convince Tyler. As they packed up for the day, their moods were still high.

Back on the drive to the hotel, Tyler remained quiet, staring out the window as the city lights passed.

His mind wasn’t idle, as he was already planning the next step.

If the rest of the NVMe units could be completed tomorrow, he would finally be able to move forward with the motherboard.

He would need to modify a motherboard to accommodate ten Valkyrie-X GPUs, four memory DRAM towers, and twelve 16TB DRAM-based NVMe units—all on a single integrated grid.

Yes, it was beyond absurd and bordering on madness, but it was absolutely necessary.

The average 2010 motherboard was built to support a single GPU, two to four DIMMs of DDR3 RAM, and two SATA SSDs or hard drives.

Even high-end enthusiast boards maxed out at three GPUs via PCIe and topped out at 24–48GB of RAM.

But Tyler needed something else.

He needed to remake the motherboard possibly from scratch into something that could survive a 10,000W power draw without combustion.

Something with heat lanes instead of just heat sinks. Something where every GPU, RAM tower, and storage module plugged into a dedicated chamber engineered for stability, balance, and thermal equalization.

It was going to be a lot of work but he will definitely get it done.

Search the lightnovelworld.cc website on Google to access chapters of novels early and in the highest quality.