Returning to his senior year of high school, he went all in and made billions by buying at the botto

March 2nd.

5:17 a.m.

The test platform was powered on again.

The control chip completes the startup.

The computing, caching, and storage modules are connected in sequence.

The scheduling system begins executing the first group of parallel tasks.

ten minutes.

half an hour.

One hour.

The yellow warning line never appeared on the screen again.

After a six-hour full-load test, the first sample chip was still running stably.

The second sample subsequently completed a 24-hour continuous pressure test.

Everyone in the lab was suppressing their excitement; several days of disappointment had made them all cautious.

Wu Huaqiang only signed his name after the 72-hour stability test, during which the package underwent twelve rounds of hot and cold cycles and all storage channels remained connected.

"The first round of engineering verification has been passed."

After he finished writing that sentence, his hand holding the pen remained on the paper for a long time.

The next second, applause rang out from behind the testing area.

The news quickly spread throughout the entire floor.

Gu Shaoting slowly took off his glasses, turned around, and quietly wiped his eyes.

He has been trudging through the mud and thorns of this self-reliant path for far too many years.

Chen Jing stood at the edge of the crowd.

Her hair was somewhat disheveled, and she had noticeable bluish tint under her eyes. The cuffs of her white lab coat still bore traces of the grinding process.

Wu Huaqiang pushed through the bustling crowd and handed her the heavy failure analysis report with both hands.

"You're listed first as the author of the failure analysis report."

Chen Jing paused, then said, "Mr. Wu, this is the result of the entire team working through the night to investigate..."

"Without your pull that logical numbering back to physical reality, we would all still be stuck in a dead end, crashing and bleeding in the process of redrawing the map."

Wu Huaqiang's eyes were resolute: "When it comes to cutting-edge technology, there's no need to be modest about who gets the credit. What belongs to someone should belong to them."

The engineers around him looked on with genuine admiration.

From that day on, no one regarded Chen Jing as Chen Mo's older sister anymore.

She has secured a unique position within the core team of Qingjia.

……

After the engineering sample passed the stability test, the actual performance verification began.

Qingjia did not use one or two of its own advantageous projects for external promotion.

Gu Shaoting brought in testing teams from two national-level laboratories in China and also asked the National Integrated Circuit Industry Investment Fund to provide complete data on internationally advanced chips and server accelerator cards of the same period.

The tests are divided into six categories: general computing, parallel computing, data retrieval, storage throughput, communication latency, and unit power consumption.

All environments are uniform.

The compiler, workload, and power consumption measurement methods are jointly confirmed by the three parties.

The results on the first day were not outstanding.

In general tasks that rely on single-core performance and a mature software ecosystem, the Qingjia sample is significantly lagging behind.

Some programs cannot even run directly and need to be re-adapted through an intermediate scheduling layer.

The engineers at Qingjia were not disappointed.

From the very beginning of its design, this chip was not intended to rely on a single core to accomplish all tasks.

The next day, the test moved to large-scale parallel computing.

The gap began to narrow rapidly.

When the dataset exceeds the local cache of a single chip, distributed storage chips and unified scheduling begin to play a role.

Traditional systems require constant data transfer between the processor and external storage.

The Qingjia sample allows computational tasks to be allocated directly on the chip closer to the data.

The morning of the third day.

The first result that surpassed the control group appeared.

In large-scale matrix computing, under the same power consumption, the continuous throughput of the Qingjia sample chip is 3.1% higher than that of the top international server acceleration chips of the same period.

As the data volume continued to expand, the lead reached 6.8 percent in the afternoon.

For storage-intensive data retrieval, it leads by 9.4%.

The effective bandwidth in the multi-core collaborative state is 12.7 percent higher.

In terms of overall target workload power consumption performance, it leads by 7.6%.

Some tests showed a lead of only a few percentage points.

Some general tasks are still more than 20 percent behind.

However, the ChipSemiconductor chip placed on the national-level testing platform uses a mature 28-nanometer process that is already in stable mass production and has a relatively controllable supply in China.

The control group consists of a 16-nanometer flagship product recently launched by an international manufacturer using 193-nanometer immersion lithography and multi-patterning processes.

With its mature manufacturing process, it forcefully crossed the gap of two generations of physical processes in a specific key area.

Inside the testing hall, several white-haired national-level evaluation experts stroked the cold data curves, speechless for a long time.

The significance of these few percentage points transcends the commercial aspect of winning or losing.

It heralds the breaking of a completely new technological path.

In the long, dark night when the extreme limits of lithography machines were completely blocked, the domestic semiconductor industry finally found a way to break through without being completely shut down through advanced architecture innovation, in-memory computing reconstruction, and microsystem packaging collaboration.

It cannot replace the need for basic manufacturing upgrades, but it has secured a precious strategic breathing space for the nation's high-performance computing infrastructure.

In the early hours of the morning, all the evaluation reports were packaged into a booklet.

Looking at the test report, Gu Shaoting suddenly asked, "The project codename hasn't been decided yet, has it?"

Wu Huaqiang shook his head.

Previously, all documents used only a single security number.

All eyes gradually fell on Chen Mo.

He was the one who first proposed this architecture.

Chen Mo turned to look out the window.

At the edge of Kyoto's deep blue skyline, the heavy night had not yet faded, and the city was still asleep.

But on the far horizon, a cold white ray of light was breaking through the dawn.

"Dawn".

Chen Mo spoke softly: "The first-generation engineering verification chip will be named Dawn One."

Gu Shaoting's gaze flickered slightly as he softly savored the two words, his heart surging with emotion: "Dawn... the first ray of light piercing the long night. A good name."

Wu Huaqiang picked up his pen and, stroke by stroke, wrote on the hard cover with the highest level of security:

【Dawn-01】

……

March 11th.

A 187-page document titled "Overview of Topology and Heterogeneous Verification of Project Dawn-1," classified as "Top Secret," was sent to the core decision-making level of the nation via a top-secret dedicated line.

Two national-level laboratories subsequently conducted retesting of key projects.

The second and third test records yielded the same results within the error range.

The message was then sent to a higher level.

In a conference room on the western outskirts of Kyoto, more than a dozen people in charge of related fields canceled their original schedules at the last minute.

The oval table had no extra decorations in the center, except for a high-strength anti-static sealed box.

The engineering sample inside the box has no commercial logos affixed; only a set of minimalist codes is laser-etched on its silver-gray heatsink surface.

QJ-CX-01

Gu Shaoting personally gave a two-hour report.

From physical defects to topology reconstruction, from packaging material failures to breakthroughs in in-memory bandwidth, there are no embellishments of advantages, nor any concealment of the severe challenges of a weak ecosystem and high yield costs.

At the end of the report, a leader in a dark jacket turned to the last page and asked solemnly, "Comrade Shaoting, does this mean that in future strategic planning, the pace of advancing to higher positions can be appropriately slowed down?"

"Absolutely impossible, there is absolutely no room for wishful thinking."

Gu Shaoting stated unequivocally: "Advanced lithography is the physical foundation. No matter how difficult or how large the investment, the hard battle of lithography machines and basic materials cannot stop for even a day."

"What Dawn proves is that even when the manufacturing process is temporarily lagging behind, we can still leverage our existing manufacturing capabilities to the fullest extent using architecture, storage, and packaging."

"It is a parallel route."

"This is also the first window that China may have secured in some areas of high-performance computing."

A long silence fell over the meeting room.

A moment later, another supervisor flipped back to the first page of the report, his fingertip landing on the signature column of the project's chief architect.

Two words were clearly written there: Chen Mo.

Those who knew the inside story were not surprised.

For the first time, many people associated this name with Star Technology, Taichu Capital, and Chuangsheng, which have recently come into view.

A moment later, the leader sitting in the main seat closed the report.

"Test data will remain sealed."

"The confidentiality level of Project Dawn has been raised, and the core team, OEM data, and key supply chain have all been included in special protection."

"Qingjia immediately prepared a second round of verification and small-batch trial production plan."

"Establish alternative lists for intellectual property rights, design software, and the supply chain of packaging materials and equipment."

"No entity may release performance data to the public without approval."

He glanced at the sample inside the protective box.

"At the same time, a special report will be prepared and submitted to higher authorities."

"Such a chip deserves serious attention from the nation."

On the day the meeting ended, Qingjia received a new official document.

Dawn-1 has been included in the special pre-research sequence of the National Integrated Circuit Major Project.

The initial performance data will be archived for ten years.

The safety, funding, equipment, and subsequent trial production of the R&D team will be guaranteed separately.

No celebration was held at Qingjia Experimental Building.

There has been no news from the outside world.

The engineering sample, which had undergone failure, slicing, repackaging, and 72 hours of continuous testing, still lay quietly in the constant temperature protective cabinet.

Through the transparent lid, you can see several newly engraved characters on the surface of the package.

QJ-CX-01.

Dawn One.

The winter for the semiconductor industry remains long, and the ice of blockade has yet to melt.

But the pioneers of technology on this ancient land already knew that once the first ray of light had torn through the long night, the bursting forth of daylight could no longer be stopped.