The world's richest man started by developing gutter oil

Chapter 298 Chapter 288: Glue CPU?

After all the employees who came here had completed the entry procedures, Chang Huazhang took these employees to Deep Blue Semiconductor. The employees responsible for auxiliary work will work in the office building of Deep Blue Semiconductor in the future.

"Director Zheng, your R&D department works in the second phase. I will take you to the second phase now." Chang Huazhang said to Zheng Weibing, who had not made arrangements yet.

"Okay, Assistant Chang." Zheng Weibing nodded.

Then more than 100 people took the sightseeing bus in the park to Building 1 of Phase 2. Deep Blue Semiconductor R&D Department was arranged in Building 1 of Phase 2. Fortunately, there were enough sightseeing buses in the park, otherwise these more than 100 people would It takes an unknown number of trips back and forth to transport them all.

At this time, Rong Zhengli and a dozen employees were already waiting for Zheng Weibing and others downstairs in Building 2.

"Dear new colleagues, this is Rong Zhengli, the general manager of our Deep Blue Semiconductor. In addition to being the general manager of Deep Blue Semiconductor, Mr. Rong also serves as the R&D director of Deep Blue Semiconductor. Because the confidentiality level of the R&D department is relatively high, I did not enter the R&D department. With the qualifications, Mr. Rong and everyone will enter the R&D department next. Mr. Rong is the R&D Director Zheng Weibing from Wanwan," Chang Huazhang introduced to everyone.

"Director Zheng, welcome to Deep Blue Semiconductor." Rong Zhengli extended his hand and smiled.

"Hello, Mr. Rong." Zheng Weibing shook hands with Rong Zhengli.

Rong Zhengli and Zheng Weibing then made arrangements for these new employees. Ordinary employees were led by Rong Zhengli's subordinates to Deep Blue Semiconductor, while Rong Zhengli personally led Zheng Weibing to Deep Blue Semiconductor.

According to Rong Zhengli's arrangement, he first took them to visit Deep Blue Semiconductor's office, and then led these people to work in their respective jobs.

"Mr. Zheng, because our R&D department involves the top secrets of the entire group, the security level of our R&D department is higher than that of other departments. Every door of the R&D department requires face authentication to pass. Staff at different levels can only Enter different work areas," Rong Zhengli introduced as he led Zheng Weibing.

Zheng Weibing didn't expect the security of the Deep Blue Group to be so strict. The level of strictness was unmatched by anything he had ever seen or heard of.

After a while, the two came to the R&D office. Zheng Weibing found that the office here was made into cubicles. There was no public area here. Each team had its own team office.

Rong Zhengli also explained this situation to Zheng Weibing.

"Each team in our R&D department has its own tasks. In addition, each team's R&D tasks are confidential, so there is little communication between the teams in the R&D department. Of course, there is no communication between the company's employees. So it feels strange.

Because the company considers the health of its employees, the company often organizes social activities, such as playing football, camping, picnics, watching movies and other activities.

Employees who take turns will interact with each other in activities organized by the company, so the relationship between employees in our company is very friendly. "Rong Zhengli introduced.

At this time, the two people had also arrived at the office of the ARM chip R&D team. The employees in the office did not stop because of the arrival of the two people. After taking a look, everyone lowered their heads and continued working.

"Mr. Rong, which company does this EDA software they use come from? Why haven't I seen this EDA software before?" Zheng Weibing asked, looking at an employee's monitor screen.

"Director Zheng, this EDA software is the Beidou Star EDA software independently developed by the group. Beidou Star EDA is more powerful than the EDA from the three major manufacturers. Beidou Star not only has the strengths of the three major EDA manufacturers, but also has functions that they do not have. To put it simply, they have We all have things they don’t have that we have.

However, this is not the most powerful thing about Beidou Star. The most powerful thing about Beidou Star is its simulation speed. Beidou Star EDA has a dedicated simulation accelerator card. The simulation speed of Beidou Star is 120 times that of the three major EDAs. I believe Director Zheng should know what this means. "Rong Zhengli said with a smile.

"120 times?" Zheng Weibing looked at Rong Zhengli dumbfounded and said.

"Yes." Rong Zhengli replied.

After getting confirmation from Rong Zhengli, Zheng Weibing swallowed. This Beidou Star EDA is really terrifying. No wonder Deep Blue's ARM chip is stronger than ARM. It takes several days for 伱ARM to simulate a chip, while Deep Blue Semiconductor only takes ten In a few minutes, your work efficiency is several times or even dozens of times higher than yours.

With the support of such efficient work efficiency, it is strange that Deep Blue cannot take off.

"Mr. Rong, what will be my next work?" Zheng Weibing asked after calming down.

"Your work will be arranged by Mr. Han himself. I will take you to Mr. Han later to see how Mr. Han arranges work for you. Mr. Han has very strong semiconductor research and development capabilities. The structure of Deep Blue 1 was designed by Mr. Han. ." Rong Zhengli smiled.

"Our boss can also design chips?" Zheng Weibing asked in surprise.

"Our boss knows a lot of things." Rong Zhengli said with a smile.

After half an hour.

Rong Zhengli took Zheng Weibing around and then took him to Han Chen's office.

At this moment, Han Chen was already making tea at the tea table and waiting for them.

"Director Zheng, this is Mr. Han, the president of our group." Rong Zhengli introduced.

Zheng Weibing looked at the young man sitting at the tea table and felt a little unbelievable, because the young man in front of him was too young.

"Mr. Han, okay." Zheng Weibing said hello quickly after regaining consciousness.

"Sit down." Han Chen smiled.

"Thank you, Mr. Han." Zheng Weibing replied.

"Have you seen the Big Dipper?" Han Chen asked after pouring tea for Zheng Weibing.

"I've seen it," Zheng Weibing replied.

"What do you think?" Han Chen asked again.

"At least three eras ahead of his peers." Zheng Weibing replied.

"Do you have the confidence to develop a CPU with performance surpassing the Core I series with Beidouxing?" Han Chen asked after taking a sip of the teacup.

"I have no confidence." Zheng Weibing replied very honestly.

He has no confidence in surpassing AMD's Pterosaur series, let alone Intel's Core.

After all, the strength of the three parties differed by more than one era, otherwise VIA would not have sold them.

"You are very self-aware." Han Chen said, taking out a document from under the table and handing it to the other party.

Han Chen took a brain capsule for this document. The document recorded the architecture description of an X86 CPU. This architecture was deduced by Han Chen based on his past life memories.

After Zheng Weibing took the document, he looked through it carefully. Zheng Weibing frowned after seeing the contents of the first few pages.

The architecture designed by Han Chen was once jokingly called glue CPU by netizens. Intel once launched a product using this architecture. This product can only be described as very bad. This product is the well-known glue CPU dual-core Intel Pentium D920.
However, after seeing the following content, Zheng Weibing showed a shocked expression, because Han Chen's glue CPU is very different from Intel's.

Han Chen's idea is to divide a whole chip into several small chips and then package them in a CPU. This idea not only requires design, but also packaging technology. Zheng Weibing doesn’t know whether existing packaging technology can achieve it.

"How?" Han Chen asked after seeing Zheng Weibing put down the document.

"If you just design a CPU, it's not very difficult. The question is, has this chip packaging technology solved it?" Zheng Weibing asked.

"The group is currently the largest shareholder of TSMC. TSMC and the group have jointly set up a laboratory. This laboratory is right above your floor. You can just communicate with them directly for the packaging process. They will produce a CPU for me within two years. Do you have the confidence? ?" Han Chen asked.

"Have confidence." Zheng Weibing replied.

"My minimum requirement is that even if it is not as good as Intel's Core series, it must be comparable to them in terms of performance." Han Chen said.

"Mr. Han, I dare not say that it will surpass Intel Core I7, but it can at least reach 80% of theirs. I guarantee that we can completely surpass them within 5 years." Zheng Weibing said confidently.

"I hope to see your finished product in two years. I have nothing to explain here. You can go back and arrange the work of the team. As long as you can achieve results, the company will definitely not treat you badly. You ask Just ask Lao Rong how much his year-end bonus was last year," Han Chen said.

"Okay, Mr. Han." Zheng Weibing nodded.

After the two arrived at the entrance of Shenlan Group.

"Mr. Rong, how much year-end bonus did you get last year?" Zheng Weibing asked curiously.

"My year-end bonus last year was more than 2000 million yuan, and the people below will get at least 20 yuan a year. As long as they can make achievements, the company will not treat anyone badly." Rong Zhengli said with a smile.

Zheng Weibing gasped after hearing about the year-end bonus of more than 2000 million. He only received a year-end bonus of several hundred thousand at Centaur. The year-end bonus given by Deep Blue Group was several times or even dozens of times that of theirs.

This was something he never dared to think about when he was at Centaur before.

September 2011, 4, China time.

Pascal Banning and his team arrived in Shenzhen City the day before yesterday. When they came, Chang Huazhang also picked them up at the airport. Han Chen also gave them a three-day holiday to let them experience the Chinese customs.

After three days of experience, Pascal Banning completely changed his view of China.

Pascal Banning discovered that what was said in the media was all lies. China was far more developed than they imagined, and China's prosperity was comparable to that of New York.

Later, Pascal Banning, under the leadership of Rong Zhengli, became aware of Beidou Star EDA, an extremely abnormal EDA software.

"Rong, what are my work arrangements going to be?" Pascal Banning asked after taking a deep breath.

"The next work arrangement for you will be arranged by Mr. Han himself. Although I am the general manager of Deep Blue Semiconductor, all work arrangements are controlled by Mr. Han. I will take you to Mr. Han's office now." Rong Zhengli said.

When Pascal Banning saw Han Chen, he had the same incredulous expression as Zheng Weibing.

He didn't expect that his new boss would be so young.

"Mr. Han, this is Pascal Banning, the chief designer of S3 Graphics." Rong Zhengli introduced.

"Pascal Banning, I'm glad you can join Deep Blue Semiconductor." Han Chen smiled.

This time Han Chen did not greet Pascal Banning at the tea table. When the two came to the office, Han Chen was sitting at the desk.

"Boss, nice to meet you." Pascal Banning greeted.

"Pascal, your next job is to develop GPUs for PC and ARM, but before that, you should read this document first." Han Chen handed a document to the other party and said.

This document records the instruction set supporting the Magic Eye engine and the architecture of a desktop GPU.

The desktop GPU architecture is the same type of architecture as Han Chen and Zheng Weibing before. They divide a whole chip into several small chips and then package them in a CPU.

Han Chen named this framework Yuanfeng Framework.

The Yuanfeng architecture adopts a unified computing unit. The unified computing unit can share resources between rendering, AI and ray tracing to more effectively utilize each transistor, thereby achieving stronger and more efficient performance than previous generation products. .

The Yuanfeng architecture also adopts a new dual-issue instruction set design, which can achieve a 2x performance improvement in execution and rendering efficiency.

As for how far the collective performance can be achieved, it depends on Pascal Banning's design.

Pascal Banning showed the same expression as Zheng Weibing after seeing the structure designed by Han Chen.

"Boss, I have never seen the instruction set in the file. No engine on the market uses these instruction sets. Can the GPU we designed be compatible with software on the market?" Pascal Banning asked.

"It is compatible with any software on the market. Game developers only need to optimize it. The reason why you have not seen this kind of instruction is because it is a supporting instruction for our independently developed engine.

The engine we independently developed is called the Magic Eye Engine. It is far superior to the Unreal Engine 3 in all aspects. This engine will be completed around July, and you can experience it for yourself by then.

The working efficiency of Magic Eye's supporting instruction set is twice that of Nvidia's CUDA.

The information you have is just an introduction. I believe I will send the information to your email later. You will understand when you open the email.

All you need to do is cooperate with Rong Zhengli in developing the GPU used by ARM. I will give you 2 years to develop the desktop GPU. I have only one requirement for the desktop GPU, which is to surpass Nvidia's products in performance. "Han Chen said.

"Boss, I will try my best." Pascal Banning said after taking a deep breath.

This time Pascal Banning did not pat his chest to assure Han Chen like Zheng Weibing did.

Because there are many unstable factors in the things Han Chen gave. Although Han Chen said that the engine independently developed by Deep Blue is very powerful, how could he dare to guarantee it before seeing the things.

"Well, okay, you can go back and arrange your team's work." Han Chen replied.

Han Chen also guessed his concerns, so he didn't force him to confirm. I believe he will be confident after the Magic Eye engine is launched.

"Okay, boss." Pascal Banning replied.

Rong Zhengli then left Han Chen's office with Pascal Banning.

(End of this chapter)