Siheyuan starts from the stationmaster

Of course, this Internet is not the same thing as the Internet of later generations. The Internet we refer to in later generations generally refers to the World Wide Web invented by Dr. Tim Berners-Lee, while the Internet at this time refers to Arpanet (ARPA) .

"ARPA" (ARPA) is the abbreviation of Advanced Research Project Agency (Advanced Research Project Agency).One of his core organizations is Information Processing (IPTO Information Processing Techniques Office), which has been focusing on research topics such as computer graphics, network communication, and supercomputers.

Arpanet was first developed and researched by the Advanced Research Agency of the Ministry of National Defense of the United States. It is the world's first operational packet switching network and it is the ancestor of the global Internet.

The term Arpanet (ARPA) is the abbreviation of the Advanced Research and Development Agency of the Ministry of Defense of the United States.It was originally just the internal communication network of the Ministry of National Defense of the United States. With the continuous development and maturity of technology, it allowed external universities, scientific research institutions, commercial organizations, etc. to access it. In 1975, it was interconnected with satellites to achieve global connectivity.

It is not easy to establish a node and connect to Arpanet in Xiangjiang, a place far away from the beautiful country and Europe. Dingtian Company spent 500 million US dollars on the equipment for the initial connection, and also invested a lot of maintenance costs every year .

The greatest convenience brought by Dingtian’s connection to Arpanet is that the company’s five R&D centers located in Cambridge, Wanwan, Silicon Valley and Xiangjiang have been integrated into one, allowing technicians to communicate in real time and communicate with each other via email .

Bai Zirui opened the email, and quickly browsed through the content of the email.This email comes from Stanford University in faraway California, and the signature at the end is John Hennessy.In recent months, Li Xuan has communicated with each other very frequently. The topics they discussed are all about a new risc architecture that Professor Hennessy is developing, which he named mips instruction set architecture.

The greatest convenience brought by Dingtian's connection to Arpanet is that the company's five R&D centers located in Cambridge, Taipei, Silicon Valley and Xiangjiang are all connected into one, and technicians can communicate in real time. mutual communication.

Bai Zirui opened the email, and quickly glanced through the content of the email.This email comes from Stanford University in faraway California, and the signature at the end is John Hennessy.

Bai Zirui has been communicating frequently with each other recently. The topic they discussed is about a new RISC architecture that Professor Hennessy is developing. Professor Hennessy named it the MIPS instruction set system architecture.

Since the mid-80s, the RISC architecture has matured technically and entered the stage of history, and the competition between the complex instruction set (CISC) and the reduced instruction set (RISC) of the processor has been very fierce.

From an objective point of view, the two have their own advantages and disadvantages, but throughout the 90s and [-]s, the academic community overwhelmingly believed that the complex instruction set (CISC) was outdated, and the reduced instruction set (RISC) was the development direction of advanced technology.

Among them are two most important leaders, namely David Patterson, a professor from the University of California, Berkeley, and John Hennessy, the president of Stanford University.

"Computer Principles" and "Computer System Architecture", two compulsory courses for computer majors in almost all universities in the beautiful country, use the textbooks co-edited by these two masters.

For a long time, a lot of content in the book is based on the MIPS architecture invented by Hennessy himself as a case, and the RISC architecture is analyzed and explained in detail, while the CISC architecture is rarely mentioned.

If the 8086 processor comes out a few years later, I believe that Intel will never risk the world's disgrace and embark on the road of CISC research.Unfortunately, when Intel was developing the 8086, the RISC architecture had not yet appeared.

In order to be compatible with the 8086 processor, the 286 and 386 developed by Intel can only continue to use the CISC instruction set, and go further and further on the road of fighting against the world.

When Bai Zirui was reborn, Intel was still extremely brilliant. Even if it fought against the whole world with its own power, Intel won a war that no one was optimistic about with the help of IBM's PC compatible machines.

Not only has it formed a monopoly in the desktop-level cpu market, but it has also counterattacked the server-level cpu market, grabbing a large share from various RISC processors.

Now that the butterfly Bai Zirui has appeared, it is questionable whether Intel can succeed as it did in its previous life.Presumably at this time, the proposal to enter the personal computer market has already been placed in front of the top executives of IBM's board of directors, and he has already carefully prepared a big meal for the blue giant.

The main course of this feast is the TRM architecture processor developed by Dingtian Company in cooperation with Xiangjiang University in the laboratory.During the turmoil of the equity incentive plan, Bai Zirui only accepted exclusive interviews from "The Economic Journal" to avoid harassment from other media.

This is because the research and development of TRM has entered the most critical moment. He put aside all other complicated things and devoted himself to the research of TRM.

Finally, a few days ago, good news came from the laboratory. The research and development of TRM has achieved initial success, and all the designs have been completed. The next step is to send it to the beautiful country for tape-out.The processor enters the tape-out stage, which means that the design is basically completed. The next step is to test the sample cpu produced by the tape-out, find out the errors and defects, and then make improvements.

It can be said that the TRM processor that Bai Zirui intends to use to compete with IBM's PC compatibles has made a breakthrough.According to normal circumstances, it usually only takes half a year from the successful tape-out to the official launch of the CPU.

In fact, Dingtian Company does not need to wait until the cpu is officially launched, as long as it confirms that the sample of the tape-out has no major defects, Aikang Computer and Dingtian Software and other branches can develop supporting hardware for the next generation of personal computers based on the relevant technical data of the TRM architecture Software development work.

When the TRM processor is officially put into production, the research and development of other hardware and software will be almost in progress. Dingtian can release its own new generation of personal computers before IBM's PC compatible machine goes on the market. caught off guard.

Bai Zirui is very clear-headed. Although the TRM processor is a RISC architecture processor that represents the direction of technological advancement, its early performance is actually very mediocre.According to many people in the industry in later generations, before DEC joined the ARM camp to develop the StrongARM processor, ARM was more like a microcontroller than a microprocessor.

The reason why Bai Zirui wanted to learn from the ARM architecture, which is not outstanding in performance, was to create an "original" TRM architecture with the effect of time difference. In Tian Ji's horse racing, the means of superior horses against other middle-level horses, using 32-bit TRM1 to suppress Intel's 16-bit 8086 and 80286.

When Intel's 32-bit 80386 processor was launched, Dingtian had already established numerous barriers in the personal computer market relying on its advantages in operating systems and application software.

At that time, the 80386 processor had no chance of competing fairly with the TRM processor, because no supplier on the market was willing to make great efforts to make its products compatible with Intel's new processor.