The competition between Intel and AMD seems to have been doomed from the beginning of their establishment.
in 1968, Intel company was founded, and then in 1969, AMD company began to formally open its business. The "struggle" between the two companies began. In 1971, the 44 developed by Intel opened the door to the development of microcomputers as the first microprocessor.
in 1978, Intel produced the first 16-bit microprocessor 886, and at the same time, Intel also produced the matching math coprocessor i887. These two chips used mutually compatible instruction sets. These instruction sets are collectively called x86 instruction sets, and this instruction system is still in use today.
people who have been in contact with computers for a long time must know that the early computer representation methods were defined according to the X86 instruction set, such as 286, 386 and 486. At that time, the CPU produced by each company had the same name, but the brand was different.
in the early days of microprocessor development, the X86 processor proposed by Intel was far from the present scenery. At that time, both IBM and Apple introduced microprocessor products, which were different in architecture, but the performance gap was not big. At that time, Intel's attitude towards AMD and Cyrix and other companies was very subtle. On the one hand, their products are completely compatible with Intel's products, which has a certain impact on their product sales in the market; On the other hand, Intel is also using the products of these companies to stabilize the position of X86 system.
at the early stage of the development of Intel and AMD, there was a little-known cooperative relationship between the two companies, which made great contributions to the establishment of X86 system status. With the continuous introduction of 286 and 386, especially in the era of 486, x86 system has dominated the civil microprocessor market, IBM has only stuck to its own territory in the server market, and Apple has been limited to certain professional fields to maintain its unique style.
during this period, people were very indifferent to the brand concept of processors. At that time, consumers only knew the Compaq 486 or IBM 486 they bought, and they didn't care about Intel or AMD. Intel, as the standard proposer, has always been the initiator of new products, and has maintained the position of the boss in the market share. AMD can only follow behind its competitors with complete compatibility as the standard of survival, more like a factory, and can only use low prices as a means of Russian and Japanese in competition, which is why AMD has always felt like a "cost-effective" brand, which is actually a beautification of low-priced products.
forced to change
1993, a memorable year. In this year, Intel changed the previous product naming method, and registered an independent trademark-Pentium for the product that people thought should be named 586. This move not only shocked the market, but also gave AMD a blow. It is time for AMD to take a new path.
since Pentium, Intel's propaganda offensive has been continuously strengthened. At that time, the slogan of "Intel Inside" has been deeply rooted in people's hearts. After two generations of Pentium II (Pentium 2) and Pentium III (Pentium 3), Intel has become the overlord of the microprocessor market. Cyrix, which has been fighting side by side with AMD, has no choice but to marry VIA under the strength of Intel and quit the market competition.
faced with Intel's Pentium series processors, AMD has been unable to compete with Intel in terms of performance, although it has been confronted with K5 and K6 series in products. Only by relying on low prices can it barely make a living in the low-end market and watch Intel continuously expand its market share. As a technology company, AMD finally realized that simple price can't make its products recognized by users, and owning technology is the key.
in 1999, AMD launched Athlon series processors, which won the attention of the industry and consumers in one fell swoop. AMD completely got rid of its following status and became a challenger who dared to compete with Intel. It was also in this year that Intel gave up the processor interface specification that had been used for many years, and AMD did not follow the changes of Intel for the first time, and always used the original interface specification, which marked that the competition between AMD and Intel entered the technical era.
A new beginning
From Athlon, AMD seems to have found a feeling. It is undoubtedly the most proud thing for AMD to compete with Intel in technology and take the lead in entering the G era. During the period of competing for the main frequency, not only did the opponent dare not underestimate this opponent, but also let consumers know AMD. Although the market share is still at an absolute disadvantage, in many surveys, AMD has surpassed Intel to become the most concerned CPU brand for consumers.
then AMD launched a series of technical offensives. after Intel launched Pentium 4 to distance itself from AMD in the main frequency, AMD vigorously promoted the concept of CPU efficiency, and at the same time, it also introduced the consumer's consumption habit of focusing on the main frequency, laying a good foundation for future development.
In 23, AMD first proposed the concept of 64-bit, which caught Intel off guard. At that time, 64-bit technology was limited to high-end server processor products, and 64-bit technology was implemented in the civil field, which made AMD take the initiative in the competition as a technology leader for the first time. At that time, Intel was very sure that it would take at least a few years for 64-bit technology to enter the civilian market, but one year later, it had to announce the launch of 64-bit processors in a hurry in the face of market trends.
In this 64-bit competition, AMD has an obvious advantage in both time and technology. Unfortunately, the weather is not fair. Because Microsoft delayed the launch of the operating system supporting 64-bit for one and a half years later than expected, and Intel's 64-bit microprocessor was also "just" listed at this time, AMD got a good applause, but the box office was bleak. Fortunately, AMD may have expected this, and its 64-bit technology in backwards compatibility.
Intel, which didn't take the lead in 64-bit, wrote another article on dual-core processors, and launched dual-core products one month ahead of AMD. AMD is no longer the small company that followed people at the beginning. After launching its own dual-core products, it has thrown out the debate of true and false dual-cores.
what shocked the industry even more was that at the end of June, 25, AMD resolutely took Intel to court, pointing out that its competitors monopolized the industry. Regardless of the outcome of this lawsuit for the time being, AMD's attitude has explained everything, and it no longer relies on following opponents and seizing the market at low prices. AMD now demands equality and is an opponent standing on the same field.
In the out-of-court market, AMD once again picked up the sharp weapon of price. In the past few years, due to the slow development of main frequency competition, there has been almost no significant price reduction competition between Intel and AMD. However, with the development of dual-core processing technology, the two companies and other competitors in the industry have improved their production efficiency, and the product price has once again become the main battlefield for Intel and AMD to compete for the market.
the survey of x86 processor market in the first quarter of 25 published by Mercury Research, a market research organization. The results show that Intel is still the leader in this market, accounting for 81.7% of the market, down .5% from the previous quarter, while AMD is up .3% from 16.9%. Both rivals are growing in the battle, and it seems that AMD has to go a little further.
product comparison
overview of product lines between amd and Intel
AMD's current mainstream product lines can be divided into two categories according to interface types, namely, low-end product lines based on Socket 754 interface and high-end product lines based on Socket 939 interface; According to the brand of processors, it is divided into Sempron, Athlon 64 and Opteron series, and there are also dual-core Athlon 64 X2 series, in which Sempron belongs to the low-end product line, while Athlon 64, Opteron and Athlon 64 X2 belong to the high-end product line. In this way, apart from different interface types, the processors of the same brand in AMD family also have a variety of different cores, which brings a lot of trouble to consumers. It can be said that AMD's current product line is very chaotic. Compared with AMD's complicated product line, Intel's product line can be said to be quite clear. At present, the mainstream processors of Intel adopt LGA 775 interface, which can be divided into low-end Celeron D series, mid-end Pentium 4 5xx series, high-end Pentium 4 6xx series and dual-core Pentium D series according to market positioning. Except for Pentium D processor, other processors currently sold in the market are based on Prescott core, which is mainly divided into grades according to the difference of frequency and secondary cache, which gives consumers a quite clear impression and is convenient to choose and buy. (In view of the fact that all the CPU products sold in the market have moved towards 64-bit, and the 32-bit CPU has no advantage in performance or price, the CPU we listed does not include 32-bit products. Similarly, the products of Socket A interface of AMD platform and Socket 478 interface of Intel are already on the discontinued lists of the two companies, while the products of Athlon 64 FX series of AMD, Pentium XE/EE series of Intel and server field are not easy to buy in the market, so they are not within the scope of this article. )
2. Comparison between AMD and Intel product lines
Dual-core processors can be said to be the biggest bright spot in the CPU field in 25. After all, with the development of X86 processor today, it seems that the traditional way to increase performance by increasing branch prediction unit, cache capacity and increasing frequency has been difficult. Therefore, when the single-core processor seems to come to an end, both Intel and AMD have launched their own dual-core processor solutions: Pentium D and Athlon 64 X2!
the so-called dual-core processor simply means that two processor cores are integrated on a CPU substrate and connected by a parallel bus. Dual-core is actually not a brand-new concept, but just the most basic, simplest and most easily realized type of CMP(Chip Multi Processors).
processor cooperation mechanism:
AMD Athlon 64 X2
Athlon 64 X2 is actually evolved from Athlon 64, which has two Athlon 64 cores, and adopts the design of independent cache. The two cores have their own independent cache resources at the same time, and the cooperation between the two cores of Athlon 64 X2 is closer through the "System Request Interface" (SRI for short). The SRI unit has a high-speed bus connected to two secondary caches. If the cache data of the two cores need to be synchronized, it only needs to be completed through the SRI unit. This design can not only reduce the CPU resource overhead, but also effectively utilize the memory bus resources without occupying the memory bus resources.
Pentium D
Like Athlon 64 X2, the L2 caches of the two cores of Pentium D are isolated from each other, but there is no specially designed cooperative interface, but they are simply merged in the front-end bus. The disadvantage of this design is that it needs to consume a lot of CPU cycles. That is, when the cache data of one core is changed, it must be sent to the Northbridge chip through the front-end bus, and then sent to the memory by the Northbridge chip, while the other core reads the data through the Northbridge. That is to say, Pentium D can't synchronize data inside the CPU like Athlon 64 X2, but needs to access the memory for synchronization, which consumes more time than Athlon 64 X2.
comparison of secondary cache:
secondary cache has a great influence on the processing capacity of CPU, which can be clearly reflected from the high-end and low-end products on the product line of the same company. As a data buffer, the size of secondary cache is of great significance. The larger the cache, the more data it can hold, which greatly reduces the waste of CPU resources because the speed of bus and memory can not match the processing speed of CPU.
In fact, it is proved that a larger cache means that more available data can be exchanged at one time, and it can also greatly reduce the occurrence of cache errors, speed up data access and make the overall performance higher.
At present, AMD's CPU is still relatively small in the design of the second-level cache due to the manufacturing process, with the highest at the high end only reaching 2M, and many low-end products only reaching 512K, which will have some adverse effects on data processing, especially when the amount of data processed is large. Intel, on the other hand, pays more attention to this aspect. For example, the Pentium D core has integrated a 2M L2 cache, which has great advantages in data processing. In high-end products, it even integrates a 4M L2 cache, which can be said to be N times that of AMD. The data obtained from some practical tests also show that the Intel score with larger secondary cache is much higher than that of AMD with smaller secondary cache.
comparison of memory architecture:
from Athlon 64, AMD began to adopt the design of integrating the memory controller into the CPU core. the advantage of this design is that it can shorten the data exchange cycle between CPU and memory. in the past, the design of integrating the memory controller into the north bridge chipset was changed into the CPU core, so that the CPU can directly access the memory without going through the north bridge, which effectively improves the processing efficiency and reduces the processing cost. However, this design not only improves the performance, but also brings some troubles. One is the compatibility problem, because the memory controller is integrated in the core, unlike the North Bridge chip.