Rebirth of the Tech Madman

...

At this time, Intel CEO Andy Grove is listening carefully to Tang Huan's vision for the next generation of personal computers, and the role that intel80486 can play in it. After all, no matter how the engineers on both sides cooperate or dispute over technical issues, the final decision will be made by the big boss.

For Intel, which has put all its treasures on the microprocessor business, there are currently two major problems-the latter has the upper hand in the dispute between cisc and risc, and the main product x86 series cpu, which has always been better than Motorola mc68000 series in performance. Competitors in the same cpu position are weaker. Even the latest generation of microprocessors from both sides—intel80486 and mc68040—continues this trend of contrast between strengths and weaknesses.

Their own products are so uncomfortable, which makes the outspoken Andy Grove want to persuade the other party to abandon Motorola's mc68000 series CPUs on the grounds that "Tang Huan owns 20% of Intel's shares," he was embarrassed to speak up.

Under this circumstance, Zheru’s profound skills in compilers, drivers, operating systems, and other low-level system software are of paramount importance. In terms of code optimization, as long as Tang Huan makes his subordinates think a little bit about any party, Can be promoted.

It has to be said that the status of software is becoming more and more important. In addition to sticking to users through the established operating habits of application software on the terminal, it also largely determines whether the new hardware can perform what the designer hopes. performance.

And Zheru has now achieved such a "supporting two with one hand", Intel has a tendency to look up to others invisibly. Of course, this embarrassment is different from being controlled by ibm at the beginning, and it is entirely due to the natural laws of ecological checks and balances.

Fortunately, Tang Huan was sure that the attitude of intel80486 seemed to surpass that of mc68040, which relieved Andy Grove a lot.

"In terms of performance, the mc68040 is indeed slightly better than the intel80486. However, the outstanding heat generation problem of the mc68040 is quite tricky. I am afraid it is difficult to make up for it through the process. This has brought a lot of influence on the design of the computer host." Tang Huan raised his hand and pointed to the outside, "You have also seen the guidance of today's conference. Multimedia and video games can squeeze the hardware specifications at any time to improve the performance contributed. In this regard, intel80486 and mc68040 still need to rely on the cooperation of philosophers. Processor to share the pressure."

Andy Grove naturally has no way to refute, so he has no choice but to aim for the future and said: "The intel80486 has integrated the cache and floating-point arithmetic unit into the chip. The goal of next-generation products is to continue to work hard to increase the operating frequency. Learn from risc's design concepts, such as superscalar. This requires Intel, Zheru, and Fangyuan to conduct further communication on patent licensing."

At this stage, the reason why the risc processor performs faster than the cisc processor is that the industry and academia are quite pessimistic about the prospects of the latter. One of the main reasons is the parallel computing capability of the former.

Compared with serial computing, parallel computing can be divided into parallel in time and parallel in space. Time parallelism is mainly the instruction pipeline technology, and spatial parallelism is the use of multiple processing units to perform concurrent computing tasks, such as superscalar.

Instruction pipeline technology splits the computer instruction processing process into multiple steps, which are then handed over to multiple hardware processing units to execute in parallel to speed up the execution of instructions-very similar to the pipeline in a factory, and computer instructions are equivalent to products on a pipeline conveyor belt. Each hardware processing unit is equivalent to a worker beside the assembly line.

Intel has implemented a certain degree of instruction pipeline technology within the intel80486-under the most ideal conditions, the intel80486 can process an instruction in a time period, and then achieves performance that is twice as fast as the intel80386 of the same clock frequency.

The superscalar architecture can be regarded as mimd, a design scheme with multiple instructions and multiple data, which often implements instruction pipelining technology, which is naturally more complicated in mechanism.

In a microprocessor with this architecture, there will be multiple internal execution units, such as arithmetic logic units, displacement units, multipliers, etc., so that the entire cpu can be assigned multiple at the same time in one clock cycle Instructions are executed in different execution units, finally achieving instruction-level parallelism.

As early as the 1960s, the supercomputer cdc6600 of the Control Data Corporation took the lead in superscalar design, but the realization of this mechanism on the microprocessor was still the precedent of the philosopher's risc microprocessor holder in the 1980s.

Successfully realized the holder of a lot of cutting-edge technology concepts such as the superscalar architecture, and its performance is superior to the crowd. Top models help Fangyuan supercomputers become the fastest computers in the world; mid-to-high end models help Fangyuan network servers and Fangyuan personal workstations compete for market share of traditional mainframes and minicomputers; lightweight models help telecommunications switches, Internet routers and other equipment with data throughput capacity Achieving a leap; the most interesting application of a specific model is undoubtedly the cpu that became the third-generation Cassidy game console. Although it has not been officially released, the evaluation media are full of praise.

Another successful application of this parallel technology is video processors, audio processors, and other multimedia hard-coded and decoded chips. For example, the current cpu does not have the ability to soft decompress and play VCD, and must rely on a special video accelerator card.

Such technologies, even if the chip is reverse-engineered in the laboratory, cannot be completely resolved. Whoever is in the front will naturally take advantage of the patent.

At this stage, Intel is facing Motorola, AMD, National Semiconductor and other companies within cisc, and there are uncertain factors in the product prospects brought by risc. It is not an exaggeration to describe the life as walking on thin ice. In terms of patents, that is, the little brothers who can bully the amd and the like, have to bend down in front of the philosophers.

"In order to make Intel's research and development smoother, there is no problem in principle." Tang Huan said happily: "I am not clear about my pursuit of microprocessor performance. You have not been OEM for intel80286 and intel80386 over the years. To get products with higher clock frequencies."

Andy Grove asked curiously: "In this raging dispute between risc and cisc, not only did Tang fail to express his attitude, he seemed to have confidence in cisc?"

"This kind of technology disputes have their own opinions. No one can guarantee that their predictions are 100% accurate. I, who represent the interests of many teams, are not easy to talk about it. However, I do still have the future prospects of cisc. Great hope." Tang Huan replied frankly: "After all, in a complex environment, the advantages of cisc cannot be replaced by risc. The real technological development trend should be that cisc continues to absorb the advantages of risc, just like Intel The next-generation microprocessor intel80586 will introduce a superscalar architecture."

Andy Grove laughed. "This should be the most exciting news I have heard in recent months."

"However, Intel seems to lack confidence in the future of its flagship product, the x86 series of microprocessors with cisc architecture. It is still preparing to release a new product of risc-inteli960 this year?" Naturally, there are some things that cannot be concealed from Tang Huan, the biggest one. Individual shareholders.

"At present, the fact that the performance of the risc processor exceeds that of the cisc processor cannot be ignored. As for the excellent degree of product design, the x86 family of microprocessors is still not the best one in the Intel R&D project sequence. We have to make two preparations. "Andy Grove explained.

To a certain extent, Intel's inteli960 is a bonus.

When intel_iapx_432, which tried to directly support the high-level programming language by hardware, failed completely, Intel turned its search for a solution to risc. Siemens, which once tried to acquire Intel but was preempted by Tang Huan, based on its own business needs, or cooperated with Intel on a project called biin to develop a fault-tolerant multi-processor high-performance computer system. Fortunately, it is online. High-end markets such as transaction processing, national grid services, military use, and industrial control such as nuclear reactor management are taking a share.

The general model of cooperation between the two parties is that Siemens, which has strong financial resources, pays, and Intel, which specializes in technology, comes out of people—mostly from the failed intel_iapx_432 project team. In addition, in order to avoid the restrictions of sensitive laws and regulations, Siemens owns the software and documentation for the distribution of work results, and Intel owns the chip design and mutually licenses each other.

However, this high-end field is also highly competitive. The system fault-tolerant application market has tandem_computers and stratus_computer, and the parallel computing market has competitors such as sequential_computer_systems, pyramid_technology, and alliant_computer_systems. IBM and Tang Huan also participate in different forms. As a result, until now, the biin project has spent nearly 300 million U.S. dollars, but the effect is minimal. Intel doesn't care, but Siemens, which is responsible for paying the money, could not help but retreat.

The upper management can use only numbers to calculate the gains and losses, but the lower engineers are reluctant to admit that the R&D process has failed. The project leader Glenford Myers lobbied vigorously that the risc microprocessor inteli960 developed for the biin system, even as a separate product-a general-purpose cpu, is excellent enough to be marketed.

Intel’s management, who had intended to bet on both cisc and risc for insurance, agreed to this suggestion smoothly.

Andy Grove finally said: "We have evaluated that the inteli960 performs well, and it should have a good prospect in the embedded application market. If the network card, scsi controller and other products use it as the CPU, it will not only meet the performance requirements. If necessary, stability can be greatly improved."

"I believe this. After all, inteli960 can be called an overkill." Tang Huan nodded. "But I have to remind one thing that Intel gave up the traditional business of memory, except because of the pressure of market competition. , Also in order to be more focused on the microprocessor business; but now you are hesitating on cisc and risc, you are suspected of going astray, you have to think twice. When you are doing marketing, will you say the new risc Is the product faster than the x86 series of processors that are contributing to the company’s main profit? Users will definitely wonder how to choose ~www.wuxiaspot.com~ series of processors. At present, they are indeed not as good as the outstanding ones, but Don’t forget that it has a large number of applications, and many of them are ace products with a wide range of applications and a solid user base. Pure technical excellence is not the only decisive factor in winning the market. You will not Not sure, right?"

Andy Grove pondered and tentatively said: "What Tang means, Intel's product development direction should be to stick to the x86 series of processors, and on the basis of cisc, introduce risc technology and take the road of integration?"

"I really think so." Tang Huan nodded very simply, "At this stage, as a separate product, the video accelerator card will undoubtedly contribute a lot of profits, but as long as you have some long-term vision, you can judge This is only a transitional solution, and future microprocessors can fully include its functions."

"So, I suggest that Intel's next-generation microprocessor intel80586, in addition to introducing superscalar, should also enrich the instruction set to directly support multimedia operations. Zheru will be friendly and open on related patents."

Andy Grove touched the hearing aid on his head, thought for a long time, and finally started to see the light in the muddy water of the dispute between cisc and risc.

After the momentous opponent left with joy, Tang Huan ushered in another wave of visitors-Sony president Akio Morita and president Norio Ohga.

The recent intensive inspections of these two men in the United States have attracted the attention of the richest man, which made him speculate that Sony is also likely to be wandering in muddy waters in a new field for him.