The shrewd Gates is the best at climbing up the pole and stealing the master, he quickly put down the unnecessary entanglement, and said tentatively: "I don't know how powerful the fourth-generation Fangyuan personal computer with the 32-bit architecture will be, if only I could see the prototype right away." ”

"You still care about how Xenix passes the POSI certification, I just give a general strategy here, and the actual implementation still requires you to cooperate with the specific responsible department." Tang Huan diverted the topic very tightly.

Although the concept of the 32-bit architecture of the fourth-generation Fangyuan personal computer was released early for the media to hype, the detailed progress is a trade secret. Including partners, no matter what the origin must be kept silent. How could Tang Huan reveal it to Gates in vain? The other party simply can't afford to pay this price, and it would be similar if it existed like AT&T.

It's a pity that AT&T has pulled a lot of chips and has a lot of chips in Tang Huan's place, but the Western Electric Company under his command, compared with Bell Labs, which is constantly launching new commercial versions of UNIX, has achieved lackluster results, which disappointed Tang Huan, who was quite hopeful that it could become an important new force in promoting Fangyuan's technical standards.

This may have something to do with the dizzying changes in the situation in the personal computer industry. You must know that almost every month, there are manufacturers in the market who withdraw gloomily, or even if they retain their brands, they also change their lineage and become a member of the two standard camps of Fangyuan personal computer and IBM-PC.

The Fangyuan workstation and Fangyuan server, which have a direct inheritance relationship with Fangyuan personal computer in the technical system, seem to be faster than the minicomputer, which makes large enterprises with traditional business thinking very uncomfortable.

In particular, the North American home video game console market collapsed last year. Personal computers play a role in this. In the end, only cheap computer maker Commodore and video game maker EA remained. It really scared the onlookers.

In short, Tang Huan took COMDEX as the symbol and launched a batch of new products every six months, which virtually made the competition in the microcomputer market extremely fierce, and many predators who wanted to come in and get a piece of the pie could only sigh with joy, and even the American Telephone and Telegraph Company was affected by this "negative factor".

In Tang Huan's license structure, although there is a concept of "priority" for collaborators, it will never spread too much to the roadmap of the "future". That is, unlike when Intel and AMD and other companies shared and cooperated, there was no clear demarcation line drawn on the possible future development of X86 family processors.

In other words, the chips that AT&T got from Tang Huan have the timeliness of not being used now, expired and invalidated, and if it is so slow in the research and development progress of microcomputers, the cards in its hand will be completely outdated.

Tang Huan was too lazy to pay too much attention to the American Telephone and Telegraph Company, after all, the other party still made a lot of money in the communications business, and the pace of entering the computer industry was careless, which was also the proper meaning of the question. If he has that worrying effort, he might as well pay attention to the research and development progress of the fourth generation of Fangyuan personal computers.

Industry observers are generally cautious about Tang's plans to aggressively promote the 32-bit architecture of personal computers. The reason for this is that the concept of 32-bit is still taken for granted as a "high-end" application, which should be considered for mainframes and small computers.

I have to say, this is a matter of vision. Although they are all 32-bit computers, they can use technical means such as code parallelism and cache acceleration to increase data processing capabilities, and distinguish products from high-end and low-end, commercial and personal use.

Of course, bringing a 32-bit architecture to a PC is a lot of difficulty and a lot of work.

Plug-and-play hardware devices, upgraded 32-bit bus, enhanced video processing performance, DMA...... The realization of such goals is a great test of managers' mastery skills in complex projects, and the slightest mistake will be like Steve Jobs developing MAC, which will drag on for several years, which will not only waste resources, but also bring a heavy burden to the company, and may also be depressed.

Fortunately, there is a help that can significantly speed up the progress of the project, that is, many technologies have been implemented on Fangyuan workstations and Fangyuan servers, which are borrowed from high-end to low-end, which is far better than groping from scratch.

A little troublesome is reflected in the supply of CPUs, Fangyuan personal computers have always insisted on walking on two legs, that is, the product line is divided into two series - Motorola MC68000 models for the mid-to-high-end and Intel X86 models for the low-end market.

But right now, Intel can't come up with a 32-bit CPU. In order to counter Motorola's MC68000, it put its hopes on the more advanced InteliAPX432 microprocessor project, and brought in Tang Huan and Siemens to help.

However, judging from the information summarized so far, the structure of InteliAPX432 is too complex, and the operating efficiency of the microprocessor manufactured by the existing semiconductor process level is significantly behind that of Intel80286, not to mention the MC68000 to Motorola.

Perhaps, the only solution is to modify the existing thinking of the InteliAPX432 and design and manufacture it in a RISC manner. But this still can't help Intel launch 32-bit CPUs as soon as possible.

And the old rival Motorola has launched a fully 32-bit architecture MC68020, and it has been successfully taped out.

Such R&D progress, of course, benefited from Tang Huan's promotion. Similar to IBM, which can modify the internal structure of Intel80286, the MC68000 license he got from Motorola is also very lenient, and there is a lot of room for processor adjustment.

MC68020 is the first true 32-bit CPU in the MC68000 microprocessor family. The previous product only had an internal scratchpad, an arithmetic unit and other structures that were 32 bits, and the external data bus and address bus were not 32 bits.

After all, the 16MB memory space that the previous 24-bit address bus can support is enough even on high-end Fangyuan workstations and Fangyuan servers.

With the continuous reduction of memory prices, computers in the high-end field. This greatly increases the possibility of configuring more than 16MB of memory.

The four-generation Fangyuan personal computer prototype in front of Tang Huan's eyes. The CPU inside is such a MC68020.

On the surface. This computer can already run smoothly, but in fact, there are still some details that need to be perfected in terms of video, audio, networking, and so on.

Tang Huan looked at the documents in his hand, and observed the actual operation of the prototype one by one according to each marked milestone.

"Not bad." Tang Huan nodded with satisfaction, "According to this progress, it can be around the time of the COMDEX conference this fall." Expand the scope of testing to the circles of the computer professional media and let them continue to help us build momentum. ”

John Chambers reminds him that "I'm a little worried that if the concept of the 32-bit fourth-generation Fangyuan PC is too successful, it may affect the existing sales performance of the third-generation Fangyuan PC and the previous models." ”

"Are you afraid that the failure of the Osborne effect will be repeated here with us?" Tang Huan laughed.

Due to the fierce competition from companies such as KayproComputer in the portable PC market, Osborne1 has only a first-mover advantage, and Osborne1, which has been surpassed in terms of quality and technology, is becoming more and more dangerous.

So Adam? Osborne is trying to regain consumer interest by releasing a series of new product plans. But not only did the goal not be achieved, but the existing users canceled the order one after another. Enter a wait-and-see state.

In the end, the poorly managed Osborne Computer Company had to go into bankruptcy protection and survive by switching to the Fangyuan personal computer camp and manufacturing low-end portable second-generation Fangyuan personal computers.

This commercial failure has become an analysis case for many business schools in the United States, and is even called the "Osborne effect".

However, Tang Huan obviously didn't think he would repeat the mistakes of the past, he had his own ideas.

"Why Adam? After Osborne released a new product plan, users entered a wait-and-see state? Not yet the first portable personal computer, the Osborne1, after the novelty, consumers gradually lost confidence in Osborne's technical prowess, and we at Fangyuan Computer Company do not have this distrust problem. ”

"More importantly, Fangyuan personal computers from the first generation to the second generation, the second generation to the third generation, every large-scale product upgrade process, the continuity has been perfected. Most of the existing software can be easily transitioned to the new hardware platform and enjoy more powerful new features with direct or minimal modifications. ”

"The user's existing investment is respected and preserved, and they naturally do not worry too much about buying the existing product, which will be a loss. After all, when a new product is first launched, even though it is very good, the price must be high, and the work cannot be delayed, so it is not as affordable as buying the current mainstream products. ”

Speaking of this, Tang Huan smiled, "Of course, the impact you are worried about is also valid, but it is also an excellent test for our marketing team. ”

At this point, the application software demonstrates the functionality of the database EBase, and a natural language AI responder called Q&A starts running.

Looking at Symantec-Symantec in the list of cooperative companies in the Copyright Notices Department, Tang Huan couldn't help but be slightly stunned.

"This program has a good prospect, and it can allow our Fangyuan computers to be placed in libraries, banks, airports, hotels and other places to provide people with simple consulting services." John Chambers nodded in affirmation.

"Gary Hendricks, the founder of Symantec, has made a name for himself in the field of natural language AI response, and the start-up capital for the company was funded by the National Science Foundation. It's just that the management is not good, and if it weren't for the cooperation we gave, it would have gone out of business. By the way, Symantec's location is the same as Fangyuan Computer, and it is also Mountain View. Charles Simeone said.

The National Science Foundation (NSF) is an independent agency of the U.S. government that supports basic research and education in science and engineering beyond the medical field. It funds 20% of U.S. federal government-funded basic research projects at U.S. universities. NSF is a major federal sponsor in math, computer science, economics, social sciences, and other fields, in particular.

Gary Hendricks's artificial intelligence project, which can be seen and funded by NSF, should have real materials, why is there no money to be made in the end? Tang Huan couldn't help but ask curiously.

"Symantec developed artificial intelligence on the platform of the DEC minicomputer PDP-11, but after the results were available, it was found that the PDP-11 no longer has any commendable advantages compared to personal computers. After a long period of no stable business, Gary Hendricks had no funds available even if he wanted to migrate to the PC platform. ”

After Charles Simeone answered, he began to sigh again, "The industry change brought by personal computers is really big, I remember when I was still working at PARC, how popular PDP-11 was, I didn't expect it to be eliminated so soon." ”

PDP-11 was indeed all the rage in the 1970s, and Bell Labs' operating system, Unix, first ran on PDP-7 and was written in assembly language. When PDP-11 appeared, Dennis Ritchie and Ken Thompson set out to port Unix to PDP-11, a new platform that gave birth to C, which supported data types.

This shows how popular PDP-11 is. It's just that the rapid rise of personal computers has directly brought this kind of small computer off the altar.

The specific implementation of the hardware architecture of the minicomputer and the personal computer varies greatly, in the case of the processors used, the general-purpose CPUs of the two major factions of Motorola and Intel are popular on the personal computer platform, while the entire processor of the PDP-11 contains four large integrated circuit chips made by Western Digital, which was famous for its hard disks in the original time and space.

In terms of memory utilization, the PDP-11 shortcomings of the 16-bit architecture are even more obvious, all applications are limited to a 16-bit virtual address space, so only 64KB of memory can be used, which is directly thrown out of the 640KB user program memory space of IBM-PC, and it is even more incomparable with Fangyuan personal computers.

Symantec has now transplanted its natural language AI response program to the Fangyuan computer platform, especially EBase as the back-end database, which is definitely time to run and hug the thighs.

"This AI does look like it works." Tang Huan also gave affirmation, "Invite the main people of Symantec Company to come over and let me see it." ”

? (To be continued.) )