Rekindle 2003
Chapter 1055: Start all over again
Chapter 1055: Starting all over again
'Child's play'?
Qing Yun shook his head upon hearing this, raised his finger, and spoke with a hint of seriousness in his tone.
"Teacher, you forgot that this is a business activity, not just a technical research."
In fact, if they were not the master and apprentice of each other, and if he did not admire the deeds of these two teachers so much, if anyone else said this today, Qing Yun would definitely retort, who the hell is kidding? !
His straightforwardness made Li Guangnan and Huang Lingyi feel a little embarrassed, and they showed unnatural expressions on their faces.
Qing Yun continued, "Teacher, don't blame me for being harsh. I can guarantee that neither you nor Senior Brother Hu Weiwu know when the Loongson chip will be commercialized!
You have no such concept in your minds at all. You think that as long as you make it, there will be no shortage of buyers.
But, really?
He turned to Huang Lingyi, with a hint of inquiry in his eyes, "Teacher, what is the R&D cost of Loongson 2B?"
Huang Lingyi blinked, unable to answer for a moment.
She doesn't take care of these things; her disciples are in charge of them.
Seeing this, Qing Yun immediately took over the conversation, revealing a kind of understanding in his voice, "I checked, the entire research and development cost of Loongson 2718 is million."
The unquestionable certainty in the words of this cheap disciple made Huang Lingyi feel a little amused.
Who is the person in charge of the Loongson project?
She sighed in her heart, well, this guy is an entrepreneur after all, and deals with money.
This cheap disciple knew everything clearly, even though the project leader didn't know.
However, she was proud enough to have spent only over 2700 million to develop a chip.
The cost of developing a chip consists of two parts: one is the R&D investment cost, and the other is the patent and technology purchase cost.
The cost of R&D investment, that is, the human and material resources spent on a chip R&D project, ranges from tens of millions to hundreds of millions of US dollars.
The cost of purchasing patents and technologies is because the chip industry is highly patented, and the research and development of chips requires the purchase of a large number of patents and technologies, and the cost is usually between millions and tens of millions of US dollars.
In other words, the sum of the two is at least tens of millions, and it’s in US dollars.
A hint of pride flashed in Huang Lingyi's eyes. She really spent every penny as if it were a dollar!
She deserves the country and the people!
But at this time, Qing Yun shrugged his shoulders and said lightly: "Teacher Huang, now that you have my funding, you can move forward faster in your research and development.
Don't always think about saving money. What we need now is time and rapid iteration, rather than small steps.
And most importantly, you have never thought about how to recover the R&D costs.”
At this point, Qing Yun saw that Huang Lingyi and Li Guangnan were stunned. He sighed with a hint of seriousness in his tone.
"Teachers, your problem now is that you are used to living in poverty and always want to save money. Now I don't need you to help me save money at all. What I need is time!
Second, you don’t have a commercial mindset.”
At this point, Qing Yun looked at Huang Lingyi with a complicated expression, "Teacher, is it true that Loongson 2B cannot be used?"
He respected this old gentleman who was truly a patriot from the bottom of his heart.
but……
He shook his head, his eyes full of regret, "Teacher, Loongson 2B can be used in industrial computers that do not require such high frequencies.
There are a lot of industrial computer market segments, including power systems, water treatment, natural gas transmission, oil exploration, traffic signals, air traffic control, monitoring, military industry, bank ATMs...
There is no blockade of X86+Windows here!
Although each market is indeed small, at least the costs can be recovered.
Moreover, this is the market that Loongson should occupy first, not to mention that the sum of these market segments is not small at all!
Do you really need to beg the Institute of Computer Science for money? The Institute of Computer Science only has a small amount of money. If you develop 2C, people might even say you are wasting money! "
Qing Yun became angrier as he spoke.
In the past, Loongson did not think about commercialization until 2007, and only began to try commercialization in 2010, and officially started commercial operation in 2018...
In the early stage, its R&D funds were completely dependent on the scientific research funds of the Institute of Computer Science, and in the later stage, they were entirely dependent on subsidized orders.
Although it was affected by the overall environment, it did waste too much time.
Qing Yun didn't want to wait any longer, so he just showed his true colors and said to Huang Lingyi,
"Teacher, apart from the fields I just mentioned, the educational machine is the best training ground for Loongson.
All the team members will come from the Institute of Computer Science and Technology. I will set up a company specifically to operate Loongson.
If HUST feels embarrassed, they can form a joint venture, I will pay for it, the company can be called HUST Loongson, and HUST can hold the majority of the equity, but I don’t want this name.”
It is pure fantasy to think about the application of Loongson in the educational machine market with an iterative mindset.
Because there will be no feedback unless it is a product accident.
But if we think in terms of recovering R&D costs, there is a whole new world ahead.
There is no better environment for business applications.
In fact, it is the same principle as Qingyun's introduction of domestic software into universities.
Cultivate the application population.
As for the famous ones...
Emperor Yun said that he had too many names, to the point where he was about to bring disaster upon himself.
A desert turned into an oasis, which would be enough even if he shared it with the Qin family. Moreover, there would definitely be a title of "XXX of Chinese Semiconductors" waiting for him in the future.
A bit full.
Although he is protected by someone now, it is precisely because he is protected by someone that he does not need to stack the BUFF golden body anymore.
If a computer overflows with data, it will crash; but if a person's reputation overflows, he will be left with 'no reward left but death'.
After realizing this, he no longer needed to defile himself, but what he had to do now was to give up his reputation.
However, this attitude of disregarding personal reputation for the country and the nation made Li Guangnan and Huang Lingyi look at him with admiration.
After all, she is a woman, and the more emotional Huang Lingyi's eyes were red, and she kept saying "Good boy! Good boy!"
The more rational Li Guangnan wiped his eyes and looked at Qing Yun seriously.
"Xiaoqing, if this is the case, how will the future architecture of Loongson be solved?
You should know that the biggest problem with Loongson's commercialization is its MIPS architecture.
It has nothing to do with the commercialization you have been advocating.”
Why is Loongson based on MIPS architecture?
Because, for the CISC-based X86 architecture, we have no path to obtain authorization.
Therefore, we can only choose the RISC architecture system that includes branch architectures such as ARM, MIPS, and PowerPC.
As for the relationship between MIPS and RISC, it can be compared to Pengpai OS and Android, the former is the secondary development of the latter.
In the RISC architecture system, only the MIPS architecture license does not restrict any changes to the MIPS architecture.
At the same time, the MIPS architecture offers the highest performance per square millimeter and the lowest energy consumption in today's SoC designs, which is its advantage.
Moreover, MIPS is a reduced instruction set structure. Compared with the complex instruction system computing structure used by Intel, RISC has the advantages of simpler design and shorter design cycle, and can apply more advanced technologies to develop faster next-generation processors.
This is the fundamental reason why Huang Lingyi was able to develop Loongson 1000 with only over 2700 million yuan and Loongson with only over million yuan.
save money.
However, the MIPS architecture also has two drawbacks that it cannot resolve.
One is a fundamental defect.
MIPS has had problems with its memory since the beginning, which has led to limitations in its memory and cache support. When the memory becomes larger, the CPU will not be able to support it.
At the same time, MIPS is a parallel thread, similar to Intel's Hyper-Threading.
Intel's hyper-threading is just a transitional product, and ARM's physical multi-core development direction is the mainstream.
The second drawback is the business environment.
The microprocessor (CPU) at the core of various modern electrical appliances originated from three projects of three companies in the last century: Intel, Texas Instruments and Garrett AiResearch Industrial Department. The three microprocessor pioneers launched were Intel 4004, TMS 1000 and CADC.
Since then, the rapid microprocessor revolution has begun.
In the early days of microprocessors, different manufacturers basically produced chips with different architectures, and it could be said that a hundred flowers bloomed at that time.
MIPS, which has been competing with Intel's X86 since its launch, has been known for its high performance, making it a success in the industrial computer and router markets. ARM, which is also based on a reduced instruction set, has been targeting the embedded low-power field since its birth and has gradually gained momentum.
At the same time, the sudden rise of the POWER-PC architecture made Abies, the absolute overlord of the semiconductor industry in the last century, see the crisis of market fragmentation.
Therefore, under Abbiemu's "coercion", Intel licensed the X86 architecture to several other manufacturers to produce processors.
Together with the rise of Windows and the optimization of the application environment, the X86 architecture began to take the lead and defeated other architectures in one fell swoop, thus monopolizing the desktop market.
So the bitter lesson that MIPS learned from its own experience is that it is not enough for a product to have very good performance, it is also necessary to maintain commercial sensitivity.
It is precisely because of the lack of commercial sensitivity that the commercialization process of MIPS has lagged behind.
These……
As a person who has experienced this, at this point in time, when Yun Di talks about the development path of chips in the next 20 years, people all over the world can only sit down and listen obediently.
Qing Yun said calmly, "We abandoned the MIPS secondary development architecture and returned to the original RISC-IV architecture released by Professor Dave Patterson in 1988."
Li Guangnan and Huang Lingyi's surprised voices rang out almost at the same time: "Give up MIPS?!"
Qing Yun nodded heavily and said decisively, "Overthrow the existing foundation and start all over again! Without destruction, there is no construction!
We went to purchase the license for the original RISC-IV architecture. I asked about it and found out that it was not expensive. The architecture developed by Patterson was originally an open source academic architecture.
Then, based on this architecture, we conducted secondary development and rewrote the instruction set, which is a self-developed instruction set."
Qing Yun thought that if John L. Hennessy of Stanford University had not yet developed RISC-V, he would have wanted to do it directly based on it.
The main difference in performance between RISC-IV architecture and RISC-V architecture lies in design philosophy and openness.
RISC-IV is one of the early reduced instruction set architectures. Its design goal is to improve processor performance by reducing the complexity and number of instructions, which enables RISC-IV to provide higher performance and energy efficiency in certain application scenarios.
However, RISC-IV has not formed a broad ecosystem, which limits its application and development.
In contrast, the RISC-V architecture not only inherits the design concept of RISC-IV, but also introduces more modern features, such as modular design, scalability, and support for multi-core and heterogeneous computing.
Another significant advantage of RISC-V is its open source and free nature, which means that anyone can freely use, modify and distribute RISC-V source code without paying high licensing fees.
However, no matter how good RISC-V is, it is not used now.
The RISC-V project only started at the University of California, Berkeley in 2010.
For now, we can only make do with RISC-IV.
Moreover, he believes that as long as one is willing to put in the effort and spend money, who knows who will be the first to come up with the RISC-V architecture!
The LoongArch dragon architecture that Mr. Huang Lingyi led his disciples to develop in his previous life is the final solution for Loongson.
Now, I am just helping the old man save ten years of detour compared to his previous life.
He was firm in his thoughts, but the firmness and determination in his words also made Li Guangnan and Huang Lingyi dumbfounded.
They were all people who had lived for dozens of years, and they knew very well that the attitude and words of the young man in front of them showed that he had made up his mind before he came... or even long before.
Start all over again.
What kind of passion is this?! And what kind of courage!
Li Guangnan swallowed hard and turned his gaze to Huang Lingyi beside him.
This is equivalent to completely recasting Loongson from architecture to instruction set to compilation language.
It is also a complete negation of Huang Lingyi's half-life's hard work.
At this time, Huang Lingyi, after a long silence, raised her head and smiled at Qingyun, "Okay! Xiaoqing, let's start over again!"
It’s about starting over, not starting over.
Starting over emphasizes starting over from an important, critical link or stage.
To start over again means to start again from the beginning.
There was a slight tremor in her voice, but more of it was relief and determination.
In fact, she knew very well in her heart that the plan proposed by Qingyun at this moment was a correct decision and a necessary decision.
If we want to be independent, we must solve the problem from the root.
Reset and start again.
Since Loongson 1 and 2 have solved the problem of creating something from nothing and the issue of life and death, it is time to go back to the furnace and recast them in the fire, eliminating the dross and retaining the essence!
Qingyun stood up and bowed deeply to the old man, thanking him for his support for his decision.
After all, the soul of the Loongson team is this old gentleman.
This great scientist, known as the "Mother of Chinese Chips", has devoted her entire life to the semiconductor industry of her motherland.
From diodes, triodes, microcomputers, large integrated circuits, etc., we have overcome all difficulties and created our own chips.
This counterattack and breakthrough can be described as "taking the task at a time of defeat, receiving orders in the midst of danger, persevering in the face of setbacks, and rising from unyielding spirit."
And at this moment, her decision will undoubtedly open a new chapter for Loongson.
But the old man waved his hands, then held his knees and sighed, "I officially started working in 62, and it has been 42 years now...
We swore on the playground of Huaqing University that we would work for the health of our motherland for 50 years!
There are still eight years left, I don’t know if it’s enough.”
The image of the old man in his previous life, who was in his twilight years, suddenly came to Qingyun's mind. His throat suddenly felt choked and his eyes unconsciously turned red.
More than 50 years...
Starting from 62, she worked for the motherland for 61 years and never stopped until the end of her life.
"Just continue the research and development, and leave the commercialization to me! In eight years, the war of resistance was won, let alone a small chip!"
Li Guangnan, on the other side, looked at the ambition of the young man in front of him and the enthusiasm of his companions, and at this moment, there was a burning passion in his heart.
However, he took a deep breath, exhaled slowly, and asked in a deep voice, "Xiaoqing, after the Longxin is recast, how do we solve the chipset problem?"
The new CPU architecture corresponds to a new chipset, which can also be simply understood as the most important part of the motherboard.
The relationship between the chipset and the CPU is interdependent. It is responsible for connecting the core of the computer - the CPU - with other parts of the machine and is an important component that determines the level of the motherboard.
Self-developed CPU and self-developed architecture mean that there will be no motherboards available at that time.
Emperor Yun, on the other hand, looked confident and answered Li Guangnan's question loudly, "Based on the current relationship between me and VIA, they can work with us to adapt.
They have experience in this regard, and in the future... within 5 years, Yanhuang Group will definitely be able to do it!
Otherwise, I will take off Qingyun's head and kick it like a ball!"
The words sounded nice, but Emperor Yun, who was always used to hiding, still set an ambush.
How could it take 5 years?!
It took less than 5 years for VIA’s chipset to go from its infancy to being the world’s number one, let alone this guy with such a high rank!
He got a lot of good stuff from VIA and Intel.
Expired patents and obsolete technologies are merely backward and obsolete compared to today's advanced technologies.
But for the early stages of rebuilding Loongson, these things are completely sufficient.
And we have obtained the authorization of AMD from Qinchuan Group. As long as we have a lot of people and money, there is no problem that cannot be solved.
But I don’t know why his two cheap teachers were so excited about it.
Li Guangnan took a deep breath, and his wrinkled face was filled with smiles. "Xiaoqing, you are ambitious! You..."
Before he finished speaking, the door of the office opened, and a middle-aged man walked in quickly and said loudly,
"Good! Good job! Qingyun! I'll give you the Institute of Computer Science! I only want the result! I don't care about the process!
But if you can’t do it within five years, I’ll kick your head like a ball!”
Emperor Yun, who was just about to enjoy the affection of his two teachers, looked at the newcomer with displeasure when he heard the words.
You don't have any vision when doing scientific research, do you?
However, when Qing Yun saw clearly who the person was, he was almost scared to death.
your sister!
The strength of the characters will upgrade as their identities upgrade, right?
Damn it, how did I meet this emperor!
……
(End of this chapter)
You'll Also Like
-
I traveled to the interstellar space to write a novel, and the interstellar readers asked for more u
Chapter 284 2 hours ago -
Longevity begins with taking care of the master's wife
Chapter 289 2 hours ago -
Douluo: The narration was modified and the plot began to collapse
Chapter 190 2 hours ago -
Chinese 2006 starts with the Beijing Film Academy Art Examination
Chapter 715 2 hours ago -
Abnormality comes, my right hand peels off everything
Chapter 99 2 hours ago -
Black technology: starting with space jump machine
Chapter 209 2 hours ago -
I rely on farming to transform the world of cultivation
Chapter 335 4 hours ago -
I failed to survive the robbery and was forced to go home for a blind date
Chapter 84 4 hours ago -
The Sims Evolved from Goblins
Chapter 748 4 hours ago -
I rely on talking to brush the copy
Chapter 205 4 hours ago