The full text will be sent tomorrow morning. Pen, Fun, Pavilion www.biquge.info has opened the starting point of this chapter or the q_q mobile client, when the time comes, you can press and hold the chapter name on the directory interface, and download it again.

……

At this time, Intel CEO Andy Grove was listening intently to Tang Huan's vision for the next generation of personal computers and the role that Intel80486 could play in it. After all, no matter how the engineers on both sides cooperate or argue over technical issues, the final decision is still up to the big boss.

For Intel, which puts all the treasure on the microprocessor business, there are currently two major problems - the latter has a slight upper hand in the battle between CISC and RISC, and the flagship product X86 series CPU, which is always weaker than Motorola MC68000 series CPU competitors in the same gear in terms of performance. Even the latest generation of microprocessors on both sides, Intel80486 and MC68040, continue this trend of strength and weakness.

His own products are so uncompromising, which makes Andy Grove, who is outspoken, embarrassed to lobby the other party to abandon Motorola's MC68000 series CPU on the grounds that "Tang Huan owns 20% of Intel's shares".

In this case, Zheru's profound skills in compilers, drivers, operating systems, and other underlying system software are crucial - in terms of code optimization, as long as Tang Huan asks his subordinates to pay a little attention to any party, he can be promoted to the top.

It has to be said that the status of software is becoming more and more important, in addition to sticking to users through the application software operation habits that have been developed in the terminal, it also largely determines whether the new hardware can play the performance desired by the designer.

And Zheru has now achieved such a point of "holding two companies with one hand", and Intel has a tendency to look up to others. Of course, this embarrassment is different from the original restrictions on IBM, which is completely due to the natural law of ecological checks and balances

Fortunately, Tang Huan's attitude towards Intel80486 seems to have surpassed MC68040, which makes Andy Grove feel a lot more relieved.

"In terms of performance, the MC68040 is indeed slightly better than the Intel80486. However, the outstanding calorific value of MC68040 is quite tricky, and it is difficult to make up for it through the process, which has a great impact on the design of the computer host. Tang Huan raised his hand and pointed outside, "You can also see the guidance of today's press conference, multimedia and video games can be squeezed out of hardware specifications at any time to improve the performance contributed, Intel80486 and MC68040 In this regard, they still need to use Zheru's co-processor to share the pressure." ”

Andy Grove naturally had no way to refute, so he had to aim for the future and said: "Intel80486 has integrated the cache and floating-point arithmetic unit into the chip, and the goal of the next generation of products, in addition to continuing to work hard to increase the operating frequency, also intends to borrow from the design concept of RISC, such as Superscalar - superscalar." This requires further communication between Intel, Zheru and Fangyuan on patent licensing. ”

At this stage, one of the main reasons why RISC processors perform faster than CISC processors and make the industry and academia quite pessimistic about the prospects of the latter is the parallel computing power of the former.

Compared with serial computing, parallel computing can be divided into parallel from the temporal perspective and parallel from the spatial perspective. Temporal parallelism is mainly an instruction pipeline technique, while spatial parallelism is the use of multiple processing units to perform concurrent computing tasks, such as superscalars.

Instruction pipeline technology splits the computer instruction processing process into multiple steps, and then hands over to multiple hardware processing units to be executed in parallel to speed up the execution of instructions - very similar to the assembly line in a factory, where computer instructions are equivalent to products on the conveyor belt of the assembly line, and each hardware processing unit is equivalent to the workers next to the assembly line.

Intel has implemented a certain degree of instruction pipeline technology within the Intel80486 - under the ideal conditions, the Intel80486 can process an instruction in a single time period, and then the performance is twice as fast as the Intel80386 with the same clock frequency.

Superscalar architecture can be regarded as a design solution for MIMD - multi-instruction multi-data, which often also implements instruction streamlining technology, which is naturally more complex in mechanism.

The microprocessor with this architecture will have multiple execution units inside, such as arithmetic logic units, displacement units, multipliers, etc., so that the entire CPU can be dispatched multiple instructions at the same time in a clock cycle, executed in different execution units, and finally realize instruction-level parallelism.

As early as the 1960s, the CDC6600, the supercomputer of the control data company, was the first to carry out the design of superscalar, but the implementation of this mechanism on the microprocessor was the first to enter the 1980s.

The top model helps Fangyuan supercomputer become the world's fastest computer, the mid-to-high-end model helps Fangyuan network server and Fangyuan personal workstation compete for the market share of traditional mainframe and small computer, and the lightweight model helps telecom switches, Internet routers and other devices have taken a leap forward in data throughput, and the most interesting application for certain models is the CPU that has become the third generation of Xiaobawang game consoles, which has not yet been officially released, but the review media is full of praise.

Another successful application of this parallel technology is video processors, audio processors, and other multimedia hardware codecs. For example, the current CPU does not have the ability to play VCD with soft decompression, and must rely on a special video acceleration card.

Such technologies, even if the chip is reverse-engineered in the laboratory, cannot be completely analyzed, and whoever is ahead will naturally occupy the advantage in patents.

At this stage, Intel is facing Motorola, AMD, National Semiconductor and other companies inside the CISC, and there are uncertainties about the product prospects brought by RISC externally, so it is not an exaggeration to say that life is walking on thin ice. In terms of patents, that is, the little brother who can bully AMD and the like, he has to bend down in front of Zheru.

"In order to make Intel's R&D smoother, there is no problem in principle. Tang Huan said happily: "My pursuit of microprocessor performance, you are not unclear, you have not been personally OEM INTEL80286 and INTEL80386 over the years to get high clock frequency products." ”

Andy Grove asked curiously, "In this raging RISC-CISC dispute, Don not only did not show his attitude, but also seemed to have confidence in CISC?"

"No one dares to guarantee that their predictions are 100% accurate, and I, who represent the interests of many teams, are not easy to talk about. However, I do have great hopes for the future prospects of CISC. Tang Huan replied very frankly: "After all, in a complex use environment, the advantages of CISC cannot be replaced by RISC. The real technology development trend should be that CISC continues to absorb the advantages of RISC, just like Intel's next-generation microprocessor, Intel Intel80586, will introduce superscalar architecture. ”

Andy Grove laughed, "This is probably the most exciting news I've heard this month." ”

"However, Intel seems to be a little less confident in the future of its flagship product, the CISC architecture of X86 series microprocessors, and is preparing to release a new RISC product this year-inteli960?"

As early as the 1960s, the CDC6600, the supercomputer of the control data company, was the first to design superscalars.

The full text will be sent tomorrow morning. If you have opened the starting point of this chapter or the q_q mobile client, you can press and hold the chapter name on the directory interface to download it again.

……

At this time, Intel CEO Andy Grove was listening intently to Tang Huan's vision for the next generation of personal computers and the role that Intel80486 could play in it. After all, no matter how the engineers on both sides cooperate or argue over technical issues, the final decision is still up to the big boss.

For Intel, which puts all the treasure on the microprocessor business, there are currently two major problems - the latter has a slight upper hand in the battle between CISC and RISC, and the flagship product X86 series CPU, which is always weaker than Motorola MC68000 series CPU competitors in the same gear in terms of performance. Even the latest generation of microprocessors on both sides, Intel80486 and MC68040, continue this trend of strength and weakness.

His own products are so uncompromising, which makes Andy Grove, who is outspoken, embarrassed to lobby the other party to abandon Motorola's MC68000 series CPU on the grounds that "Tang Huan owns 20% of Intel's shares".

In this case, Zheru's profound skills in compilers, drivers, operating systems, and other underlying system software are crucial - in terms of code optimization, as long as Tang Huan asks his subordinates to pay a little attention to any party, he can be promoted to the top.

It has to be said that the status of software is becoming more and more important, in addition to sticking to users through the application software operation habits that have been developed in the terminal, it also largely determines whether the new hardware can play the performance desired by the designer.

And Zheru has now achieved such a point of "holding two companies with one hand", and Intel has a tendency to look up to others. Of course, this embarrassment is different from the original restrictions on IBM, which is completely due to the natural law of ecological checks and balances

Fortunately, Tang Huan's attitude towards Intel80486 seems to have surpassed MC68040, which makes Andy Grove feel a lot more relieved.

"In terms of performance, the MC68040 is indeed slightly better than the Intel80486. However, the outstanding calorific value of MC68040 is quite tricky, and it is difficult to make up for it through the process, which has a great impact on the design of the computer host. Tang Huan raised his hand and pointed outside, "You can also see the guidance of today's press conference, multimedia and video games can be squeezed out of hardware specifications at any time to improve the performance contributed, Intel80486 and MC68040 In this regard, they still need to use Zheru's co-processor to share the pressure." ”

Andy Grove naturally had no way to refute, so he had to aim for the future and said: "Intel80486 has integrated the cache and floating-point arithmetic unit into the chip, and the goal of the next generation of products, in addition to continuing to work hard to increase the operating frequency, also intends to borrow from the design concept of RISC, such as Superscalar - superscalar." This requires further communication between Intel, Zheru and Fangyuan on patent licensing. ”

At this stage, one of the main reasons why RISC processors perform faster than CISC processors and make the industry and academia quite pessimistic about the prospects of the latter is the parallel computing power of the former.

Compared with serial computing, parallel computing can be divided into parallel from the temporal perspective and parallel from the spatial perspective. Temporal parallelism is mainly an instruction pipeline technique, while spatial parallelism is the use of multiple processing units to perform concurrent computing tasks, such as superscalars.

Instruction pipeline technology splits the computer instruction processing process into multiple steps, and then hands over to multiple hardware processing units to be executed in parallel to speed up the execution of instructions - very similar to the assembly line in a factory, where computer instructions are equivalent to products on the conveyor belt of the assembly line, and each hardware processing unit is equivalent to the workers next to the assembly line.

Intel has implemented a certain degree of instruction pipeline technology within the Intel80486 - under the ideal conditions, the Intel80486 can process an instruction in a single time period, and then the performance is twice as fast as the Intel80386 with the same clock frequency.

Superscalar architecture can be regarded as a design solution for MIMD - multi-instruction multi-data, which often also implements instruction streamlining technology, which is naturally more complex in mechanism.

The microprocessor with this architecture will have multiple execution units inside, such as arithmetic logic units, displacement units, multipliers, etc., so that the entire CPU can be dispatched multiple instructions at the same time in a clock cycle, executed in different execution units, and finally realize instruction-level parallelism. (To be continued.) )