"Not interested!" Lovo was brief, perhaps in a temper, and his words were unceremonious, as if he had no regard for House's face.

I see, I see, House muttered to himself as if he understood something. Observing the words, House saw that since Lovo was not interested at all, he surprisingly did not argue, but silently packed up the box, bowed and exited the office.

It's not that House can't speak, as a marketer, he was promoted by Intel as the first image of 386, how can he be like a rookie?

There is no gratuitous hate, and there is no gratuitous love, and as early as the early days of 386 development, Intel executives discovered that IBM seemed completely uninterested in their next-generation processors.

Grove was optimistic, thinking that maybe IBM didn't see the finished product and ignored it, so he suppressed some dissenting voices with his prestige.

It's just a rumor that spreads quietly, especially after Lovo announced that the next generation of machines will not use VS's godson system, and people are panicking, saying that IBM is developing its own system and Intel's processors are becoming obsolete.

In Grove's eyes, he naturally doesn't take this level of remarks seriously, he only knows that 386 can dominate the world as soon as it appears, and he also has his own calculations.

In the era before 286, forced by reality, Intel Corporation licensed its processor manufacturing technology to many companies, and even more was forcibly purchased by IBM 20% of the shares, forcing it to modify the circuit structure of 286.

The popularity of PCs and their compatible machines has made Intel's chips suddenly become hot, not only Intel has made a lot of money, but those authorized manufacturers have also made a fortune, and there is no need for research and development, as long as there is a license, you can manufacture the same processor as Intel 286, and even because of the high manufacturing technology, you can also create a higher frequency 286.

Due to the invasion of integrated circuit products of Japanese companies, Intel's requirements for quality at that time were unparalleled, so the frequency of 286 has not gone up, from 8 megahertz to 16 megahertz, divided into three grades.

And those companies that have licenses, such as Texas Instruments, can force the frequency to increase to 20 megahertz with superior manufacturing technology.

Although there are few such products, they are regarded as treasures by those compatible machine manufacturers, and they have always been high-priced and advertised to compete with IBM, and IBM cannot use these processors because of contract restrictions, so it is dumb and suffers a lot, which is why the above-mentioned Lovo complained about House.

IBM doesn't want to live like this, in order to make the AT machine accepted by commercial users, it has completely given up the cheap XT machine market, and now the AT is 286 has just 'improved', but it is compatible with the machine so '' how can it not be troubled, so it can only frequently put pressure on Intel, anyway, he is a shareholder of Intel.

Once in a high position, people's mentality is naturally different, Grove hated this kind of constrained days, and decided to take advantage of the 386 processor to turn the sky upside down, no longer authorize any company, and have money to earn by himself, isn't it happy?

It's just that now it seems that Lovo doesn't have a cold at all, and this time it's like a thunderbolt hitting Grove's heart on a sunny day.

Why is it pinched like this? Grove was puzzled, thinking of the treatment of 8088 and 286, these 386 seem to be raised by his stepmother, and he shouldn't be.

After several meetings, no one could say why this happened, and Grove had to ask House to continue public relations and increase his efforts, at least to find out IBM's true intentions.

House had no choice but to hold his breath and continue to lobby the Boca branch. It's just that this time it's even more unfortunate, I didn't even see Lovo's face, and he was greeted by a reasonable IBM representative, who said a lot of good things, but there was no progress, and the representative turned over and over that sentence: "You don't need 386, as long as the faster 286, there is no 32-bit software on the microcomputer, so ......"

"The way to make 286 faster is to use 386!" House took the trouble to explain the difference between them, "not only can the current 16-bit software continue to be used, but the 32-bit software 386 will also be supported in the future." ”

Ren Haus was exhausted, which meant that he did not move, nor was he irritable, but just listened with a smile, which made House feel bored.

Grove was at a loss, and since IBM was not interested for the time being, the enthusiasm of developing 386 to catch up with the schedule was also relaxed. On the one hand, he asked House to carry forward the kraft candy 'spirit' and continue to pester IBM, and on the other hand, he also asked the development department to improve the process, so that the 386 could enter the mass production stage as soon as possible.

Grove doesn't believe it, and he can't live without IBM and Intel.

Intel is extremely depressed, but the long-term leadership of the VS Research Institute has come out with good news, the graphics workstation project that has been developed for more than a year has entered the final stage, the theme hardware has been fully developed, the prototype has passed the test, and all that remains is the graphics application software developed by Dr. Warlock's team, such as pfhotoshop, which has not yet been completed.

This graphics workstation was called alpha by the project team, taking its first meaning.

The Alpha also deserves this title, and the entire machine's shell is made of stainless steel, with a brushed exterior and a 'fine' glittering throughout. The volume control is excellent, even if the basic configuration of dual processors, 2 megabytes of memory, 14-inch color display, and 120 megabyte hard disk (multi-hard disk) is only one round larger than that of ordinary microcomputers, which is more than enough on the desktop.

Although the size is small, the 'sexual' energy is not ordinary. In order to test the real "sexuality" of this machine, the madmen of VS Research Institute have searched for almost all the microcomputers and minicomputers they can buy on the market with the acquiescence for a long time, and have written all kinds of programs for testing, and also invited some well-known people in the industry to observe with reporters.

The result is that to the surprise of all non-specialists, the integer 'sex' of this alpha is far superior to that of DEC's minicomputers, and none of the machines tested in comparison are its opponents, and the floating-point 'sex' is also very 'excellent', but slightly inferior to those that specialize in 'department' with floating-point coprocessors.

These are just the basic 'sex' capabilities of a machine, and in the field of graphics processing, which Alpha is best at, the results of other machines can be said to be miserable, Alpha can be said to be not an order of magnitude rival with them at all, easily winning, excellent graphics bus and independent processing system design make its graphics processing power span an era.

Although those reporters have been in the computer industry for many years, they did not expect that this test would be this kind of result, in their minds, this little-known research institute that has just been established can have any products, at best, the software skills of VS can be optimized without being at the bottom, who would have thought that it would be such an overwhelming 'sex' victory.

Those in the industry who have been in the industry for a long time are amazed, they know a lot about this, you must know that graphics processing has always been a difficult bone for computers, and the performance of alpha has surprised them, and they are all impressed.

More skeptics said they didn't believe in the home test at VS Research Institute, and asked to take the machine apart to see if there was any cheating. The madmen in the research institute naturally happily promised to let these guys see what was going on, and they wouldn't lose their technology anyway.

The result is that there is no cheating, of course there will be, this is the 'sexual' ability that alpha should have, and it has been secretly happy for a long time.

Alpha can have such a 'fine' performance naturally has its reason, these tests are only carried out on the public platform, all the test programs are now programmed, and the test is only some benchmark 'sex' performance, in this respect Alpha will naturally not lose to any opponent.

However, 'sex' can be attributed to 'sex' ability, and it can only be sold if it is actually applied, but the application software has not yet been completed, so this long-term will naturally not be mentioned.

Moreover, the good performance of alpha is inseparable from the design of the processor.

The processor of this prototype is no longer the K32 originally ordered, but an improved version of the R&D team led by Dr. Yu, which is internally called K32-Pro.

It is said that it is an improved version, in fact, it has taken a long time, from the success of K32 to start research and development, it has been more than a year now, under the research and development of the team led by Dr. Yu, its 'sexuality' leap has also made Chang Chang quite frightened.

Adhering to the extremely 'lean' and simple design of the K32, Dr. Yu did not change much to the instruction set, except that two new instructions were added, from 32 to 34, and there were only two memory addressing methods, which greatly simplified the programming process.

In addition, due to the improvement of the process, the chip can accommodate more transistors, and the K32-Pro has also doubled the internal high-speed registers to 128, so that various data instructions can run more at high speed in the registers without the need for the CPU to issue instructions to find in memory.

Since K32 uses pipeline technology, and K32-Pro is also this design, the demand for data for high-speed processing of 'fine' instructions is also rising, and it can be said that it is a data devouring machine, which brings a problem.

The large-scale register makes the data processing faster, but the speed of the memory is far from keeping up with the speed of processing, which also makes the K32-Pro's pipeline processing become a decoration, and there is no data to wait, so that the 'sex' energy can be greatly reduced.

Dr. Yu also encountered Intel's problem, which was even more fatal, because the 386 does not use pipeline technology, so the waiting of the processor's clock cycle does not have much impact on its 'sexual' performance, while the K32-Pro is very sensitive to this.

This makes it difficult for the R&D team to ride the tiger, because the core of the k32-pro does not require microcode circuitry, so it is very 'fine' and small, and it can be said that it is the smallest processor core in the world, which only consumes about 35,000 transistors (the same is true for ARM processors).

Such a small core and some peripheral functional circuits only increase the area of the chip a little, and Dr. Yu even wants to integrate two data processing pipelines on one chip, which is completely possible, compared to the nearly 280,000 transistors used by 386.

However, after simulating it, it turned out that it was not the case at all, and the two streams seriously aggravated the time of waiting for data, which made the 'sex' energy worse than the same line.

This made it difficult for Dr. Yu to make a trade-off and put a lot of effort into solving the problem. In the end, an engineer from West Gate suggested adding cache, which might alleviate the problem of data waiting.

Cache is not big, only 2kb, not bytes, bits, 2k bits, but it occupies an area equivalent to the original entire chip, doubling the cost of k32-pro.

But it's worth it, with the addition of cache K32-Pro, the problem of data waiting exists but is already insignificant, the 'sex' ability is actually more than double that of K32, and the 8 megahertz version can process four million instructions per second, which no processor can match at the time (386 is about 400 to 5 million instructions per second, but the frequency of 386 is from 20 megahertz).

If the chip is large, it will naturally affect the yield and increase the difficulty of testing, which also makes Dr. Yu a headache, and he can't do anything after exhausting the methods in the chip factory.

I never thought that this problem would be solved by a guy who was less than twenty years old, which made Dr. Yu sigh with emotion.

This lucky guy is Stephen Zhou, who has been introducing the past for a long time, this kid is now 19 years old, and he was still studying when he was VS, and later because he was more interested in hardware circuits, he was admitted to the processor development team.

The young man was very studious, and he originally wanted to quit his job and concentrate on going to college, but after a period of contact with him, Dr. Yu admired him very much and persuaded him: "You go to school but learn some outdated things, how can you have the fun of creating here?"

Stephen Zhou thought about it, the people he came into contact with here were some *** masters, and what he could learn seemed to be no worse than the university, and he could make a considerable amount of money, so he decided to stay because he was born in poverty.

Stephen Zhou, who was learning and doing, found a problem, with the improvement of integration, both design and testing are very difficult, and it is even more difficult to find errors in a finished product.

One day, he had a whim, whether it was possible to add some circuits for testing when the processor was designed, and when it was done, he only needed to 'excited' it to simply confirm that there was no problem with the chip.

He shared this seemingly simple idea with his colleagues in the R&D team, but he was praised – no one thought of it.

Dr. Yu also thought that this was feasible, and it was rare that he was very supportive of this young man, and then he came up with a way to improve the yield rate, that is, to add redundant circuits to the chip design to replace those circuits that were damaged in manufacturing.

"That's a great idea, so you're in charge of the design. ”

Dr. Yu's gentle and clever words made this young man take the lead.