Cache

Fuxing Base was still small. Once it expanded in the future, it would be impossible to keep the base running without purchasing hundreds or even thousands of units.

Especially since Han Yang planned to begin excavating an underground base once his computing power received another upgrade.

The Revival Mall on the surface was merely a cover. For safety, Han Yang had decided from the very beginning that future development would definitely take place underground.

But Han Yang certainly could not hire a large number of workers and bring in a vast amount of heavy equipment to excavate the underground base.

He would still have to use robots controlled by his own computing power to do the digging.

By then, the demands on his computing power would be even greater. And once the base was completed, operating more equipment and conducting more research would require no small number of robots as well.

Whenever he imagined controlling thousands of robots at once as they bustled about the base after his computing power grew stronger, Han Yang's heart filled with anticipation.

But for now, he could only look forward to it. Without developing more advanced chips and increasing his computing power, everything was just a mirage.

Time passed slowly. Half a month after the first trial production run, the robots were still following their usual routine. One of them carried a box of materials piled in the corner of the laboratory to the start of the production line, opened it, and poured the materials into the feed inlet.

The stationary robots worked together with the machines to process this batch of materials. The ten mobile Multifunctional Robots moved back and forth through the production line like conscientious workers.

Feeding, operation, sorting, assembly, output—everything proceeded in perfect order.

Han Yang was the only person in the vast laboratory. At that moment, he was still sitting before the computer, listening to a mathematics lecturer. Behind him, everything was handled by machines and robots, just like a fully automated assembly line.

When this batch of raw materials emerged from the output port as cables of a specific model, were bundled together by the robots, and sent to specialized testing equipment for inspection, the figures on the instrument immediately lifted Han Yang's spirits.

After half a month of repeated revisions, adjustments, and trials, the first batch of cables that met Han Yang's requirements had finally been produced.

A total of one hundred cables were produced in this first batch. They were 12.5 centimeters long, with tiny connectors at both ends, and looked exquisitely crafted.

On the other side, another machine rumbled as it produced tiny objects no larger than grains of rice, which were likewise sent to the testing equipment.

These were all transistors.

Cables were equivalent to circuits, circuit boards were equivalent to silicon wafers, and silicon wafers, circuits, and transistors were the three most important components in constructing a CPU.

In standard CPU production, transistors and circuits were directly photolithographed onto silicon wafers, making such cumbersome methods completely unnecessary.

But Han Yang could only use this method to substitute for photolithography technology.

At that moment, the circuit boards that met Han Yang's requirements, along with the first small batch of cables and transistors, were ready.

With a thought from Han Yang, a mobile Multifunctional Robot carefully picked up the materials and placed them on the work platform.

The stationary robots on the platform were immediately infused with Han Yang's computing power.

Under Han Yang's direct control, production of the new-generation CPU finally officially began.

During this long period of study, Han Yang had gradually formulated the architecture and manufacturing method for the new-generation CPU in his mind.

Compared with the previous-generation CPU, this new-generation CPU had a more complex logical structure. Han Yang had added several self-created instruction sets and adopted a pipelined design for the first time.

Pipelined circuits greatly improved efficiency by allowing multiple instructions to perform different tasks at different locations in the circuit at the same time.

This was a highly mature technology. In modern society, even the oldest CPUs had already adopted it extensively. But for Han Yang's hand-built circuit-board CPU, this was the first time it had been used.

According to theoretical projections, this technology would greatly improve the performance of the CPUs he manufactured.

In addition, Han Yang had added a 4 KB cache to this CPU for the first time.

Modern CPUs generally had three levels of cache to maximize performance. But for Han Yang, creating even one cache structure was already the limit of his current technology.

The addition of so many new designs made this CPU architecture far more complex than before, and accordingly, much more difficult to manufacture.

Although Han Yang had designed this architecture in advance, he knew that theory was ultimately just theory.

During actual manufacturing, countless new problems he had never discovered before would inevitably emerge one after another.

Deal with soldiers with generals and floodwaters with earth; solve each problem as it appeared. That was Han Yang's only viable plan at the moment.

Sure enough, just as Han Yang had expected, problems emerged before the first batch of five hundred transistors and one hundred cables had even been used up.

The soldering precision was insufficient.

The cables and transistors were too tiny. When they were soldered onto the circuit boards, the solder joints affected one another, causing short circuits.

This was a problem that had never appeared before.

Han Yang spent considerable effort puzzling over it. In the end, he paid a high price to urgently purchase a high-precision soldering gun and specialized solder, resolving the issue.

As a result, the first batch of transistors, cables, and circuit boards were all scrapped and turned into garbage.

Fortunately, the second batch of transistors and cables had already been produced, so Han Yang did not have to keep waiting.

After solving this problem, Han Yang continued manufacturing.

Then, before long, a second problem appeared.

Han Yang discovered that his original design had a bug. Two modules within the CPU were too far apart, located on two different circuit boards, and had to be connected by a cable sixty-two centimeters long.

But cables produced with his current technology could not meet the required precision and speed.

With no other option, Han Yang could only adjust the architecture, placing the two modules on the same circuit board and moving the module originally assigned there onto another circuit board.

But this affected the overall structure once again. A single change set off a chain reaction, and the communication reliability, communication latency, and other factors between all modules had to be reconsidered.

Fortunately, this was only an adjustment to the overall structure and did not involve the details, so it was not too troublesome. Han Yang spent five days completing all the adjustments.

Afterward, manufacturing continued.

Three days later, Han Yang's work stopped once more. He began researching and thinking again. After a two-day interruption, work resumed, only to stop again after continuing for four days. Then it continued for another three days before stopping once more...