Military Industry of Great Powers: Build the Strongest Flanker.

Chapter 216 Give Linnard an offer

ARM brought five engineers this time, headed by Dr. Sylvain Monsolo, as well as architecture engineer Francisco Garland, NEON unit designer Robert Hus, cache storage engineer Philip Neville, and Branch prediction and sequencer algorithm engineer Fatemire Bajramaei has a strong lineup.

You must know that although ARM has now expanded to fifty-three engineers, when it was first established a year ago, there were only twelve engineers, and these five people were among those twelve people.

Dr. Sivaner said to Zhao Guoqing enthusiastically: "Mr. Zhao, I am very happy to meet you. You are the big boss of ARM, and you are also a big customer of ARM. In fact, you are the first big customer of ARMv6. I am very happy." It would be an honor to serve you and your company, in any capacity!”

Zhao Guoqing shook hands with Sivaner without exchanging too many pleasantries. He used fluent English to communicate with Sivaner about the specific work.

In fact, at this time, Zhao Guoqing was not acting as a boss, but rather as a Party A making unreasonable demands.

"The architecture still needs to be optimized and higher computing efficiency needs to be achieved. I hope to use pipeline design and parallel data processing in the algorithm..."

Zhao Guoqing took out a stack of paper and made a rough draft, drawing the ARMv6 processor structure. When the data processing volume increases to a certain level, there will be more than a dozen units in the entire processor. Only one unit processes data, while the other units are in a wait-and-see state. .

It's the same reason as when one side of a multi-core processor has difficulties, everyone is waiting and watching.

Sylvan soon realized that this trip to China might not be a good job. Their boss, or their client, was asking them many difficult questions, which were difficult to solve.

"Mr. Zhao, we did not expect this situation. In fact, ARM6 is designed for embedded devices. This kind of data congestion situation is difficult to encounter!"

Zhao Guoqing shrugged helplessly and said, "Now that you have met, Dr. Sylvain, as you said, you can solve all the problems we encounter."

Philip Neville said: "This is a microcomputer usage scenario, Dr. Sylvain, I think we can do it. If ARM can develop on microcomputers, it is better than just developing embedded systems!"

"..."

It took ten days to demonstrate the entire plan, and the eight-level pipeline design was determined, setting up a 128kb data instruction cache, 16 floating point units, 8 integer units, using ARM's conventional AHB-LITE bus interface, and expanding the 72-pin SIMM. Memory interface, extended IDE parallel data interface.

As for peripherals, Zhao Guoqing believes that they are compatible with the ISA bus interface launched by IBM.

Sylvain rejected it, and Sylvain said: "Mr. Zhao, the ISA bus and the AHB-LITE bus have different definitions, and there is no way they are compatible!"

Teacher Xiao Zhao’s opinion is also very firm. AHB also needs peripherals and can be compatible with ISA in the form of a transfer. I only need one interface.

"The performance will definitely be lost!"

Zhao Guoqing said: "As long as 80% of the performance is enough!"

The ISA bus is the most widely used in the field of microcomputers. For example, printers, professional equipment, faxes, and input and output devices all use the ISA bus. If it is not compatible with the ISA bus, then the microcomputer based on ARMv6 will be half disabled.

This is also the power of the first-mover advantage. As more devices are developed around the standards you set, dependence will naturally form.

The development process is tantamount to a battle. On the premise of ensuring the livelihood of ARM engineers, Teacher Xiao Zhao also carefully arranged more than 30 assistants.

These people understand simple English communication, and started learning the instruction set, structure, and working principles of each module of the ARMv6 processor as early as two months ago.

At the same time, Mr. Xiao Zhao received an offer from Motorola. Motorola was willing to transfer the DRAM business unit to Zhongxun for one dollar, including all the DRAM patents they held. The prerequisite was that Zhongxun could properly arrange their 237 DRAM employees. Memory design engineer.

Zhao Guoqing arranged for Fred to establish the China Communications Research Center in the United States.

Zhao Guoqing said: "Mr. Fred, I need a mechanism for Chinese engineers to communicate with Motorola engineers. We will all learn what they want from each other!" Fred said: "BOSS, they He is no longer an engineer from Motorola, but from now on he will be an engineer from CITIC America Research Center!"

All right! The data of 1M capacity DRAM memory particles has been sent over. Zhongxun's chip processing level is undoubtedly top-notch. They tried an -nanometer process and successfully taped out the chip in one go.

The good news doesn't stop there. Toshiba responded to Sinocom's offer to Toshiba, and they are willing to transfer the technology patents of NAND flash chips for US$5 million.

Fred said: "BOSS, Toshiba is really blackmailing. NAND flash storage technology is not worth so much money at all. Two years ago, Toshiba wanted to use NAND flash technology to compete with Intel NOR flash storage technology, but Toshiba failed and NOR flash became Regarding storage standards, no manufacturer in the world except Toshiba is willing to use NAND flash!"

In Fred's opinion, a patent that is not recognized is worthless.

Zhao Guoqing shook his head. Although Toshiba seemed to be asking for a lot of money, it was only five million US dollars. Well, it was worth it.

NAND flash currently has almost zero market share. However, according to normal history, two years later, Samsung obtained the patent authorization of Toshiba NAND flash. Based on this, it developed 3D NAND flash, and the world began to pay attention to this technology again. technology.

One of the core elements of this technology, the non-volatile memory unit, is attributed to Professor Shi Min, a Chinese scientist.

He was the guy who should have won the Nobel Prize in Physics.

Zhao Guoqing took a long breath and felt unfair for Professor Shi, but the patent was finally back.

"Send someone to sign the contract, it's only five million dollars!"

If Toshiba wants to get the patent back, it won't be a matter of US$5 million...

……

August 1991, 8, Department of Computer Science, University of Hollsinki, Finland.

Linus Torvalds quickly returned to his dormitory after class as usual. He had been doing this every day for the past three months.

Eat, go to class, and go back to the dormitory.

So much so that Linus's girlfriend, Lucy Silvis, couldn't bear the boring life and happily broke up with this science boy.

Linus didn't hold him back. In fact, he also felt that when Lucy was with him, he spent a lot less time with the computer.

"Okay, Lucy, thank you for your understanding!"

"Uh, Lynyrd, you're just a freak..."

Linnard was in a good mood. When he returned to the dormitory, he saw a stranger at the door of his dormitory. This person was obviously not his roommate.

"Okay, Mr. Linnard, I am Jiang Yaoxian from Zhongxun Open Source Development Laboratory. I want to share an interesting thing with you. Can we chat for a while?"

……

(End of this chapter)