In most cases, "A" stands for Northwood core and Pentium 4 processor of 400MHz FSB to distinguish Pentium 4 processor of Willamette core with the same frequency. Specific to the processor plane number, it can be confirmed by "5 12K/400" in the "simple number", while the corresponding Willamette is "256K/400".
"B" stands for Northwood Pentium 4 processor with 533MHz FSB, which can be distinguished from "5 12K/533" of "A" in number.
"C" is a Northwood Pentium 4 processor with 800MHz FSB, and the serial number is "5 12K/800".
"E" is the latest Pentium 4 processor based on Socket 478 and rescott core, and it is numbered "1M/800" because it has 1MB L2 cache.
Please pay attention to some exceptions. The two Prescott core processors are also marked with "A", namely 2.4A and 2.8A, which do not support hyper-threading, and both are 533MHz FSB marked with "1M/533".
By associating the suffix with the serial number in the above way, we can know that the mainstream Pentium 4 * * has "256K/400" (no suffix), "5 12k/400" (a), "512k/533" (b) and "5/kloc-0".
● Pay attention to the stepping value of Northwood.
Northwood Pentium 4, which has the largest sales volume, includes the aforementioned A/B/C series. At the same frequency, the performance from high to low is C→B→A B → A, but even if they are all "C", we should pay attention to the stepping value of the processor. Generally, Northwood has three steps: B0, C 1 and D 1. The usual practice is to choose the latter step, that is, D 1. D 1 stepper usually has multiple core voltages (Intel is gradually reducing power consumption). This kind of processor generally does not recognize the core voltage in the "simple number", but can find it by looking at the S-Spec number. Because the S-Spec number is irregular, the S-Spec value and corresponding step size of Northwood processor with mainstream frequency are listed at the end of this paper for reference.
The meaning of suffix j and E0 step
Intel claims that the suffix J indicates that the processor supports hardware antivirus function (similar to Athlon 64, it can be opened in the operating system after installing WinXP SP2). It is understood that Intel's new Prescott of E0 process should support this function. In addition, E0 program has enhanced temperature control function. However, the author noticed that not all E0 stepping Prescott processors will be labeled with the suffix J. In addition, in Socket 478 processor, we also found that the Prescott core of E0 process exists, but this kind of processor will definitely not be labeled with the suffix J. Conversely, does the processor with the suffix J support the enhanced temperature control function? Officials say they don't support it. However, I don't think it is possible for Intel to block this function artificially, so it is smarter to choose Prescott with E0 process.
How to judge whether it is E0 process? It also depends on the S-Spec value. Because we can't see the difference between E0 and other processes from cache and FSB. From the official Intel processor number list, we can find that the S-Spec of E0 stepping processor includes:
Socket 478 platform: SL7PL, SL7PK, SL7PM, SL7PN, SL7PP, SL7KD；; LGA 775 platforms: SL7PT, SL82V, SL7PR, SL85V, SL87L, SL82X, SL7PU, SL7PW, SL7PX, SL82Z, SL7PY, SL7PZ, SL833, SL84X, SL7Q2, SL7NZ, SL82U, SL84Y and SL72P.
Prescott is a mess, be careful!
Pentium 4 with suffix f and suffix p
The suffix F stands for supporting EM64T, which is a 64-bit extension of Intel. With the help of S-Spec number, it can be found that there are models in D0 that support EMT64 step by step from Pentium 4. In step D0, SL7LA, SL7L8 and SL7L9 can support EM64T, that is, Pentium 4 F. In step E0, SL7PX, SL7PZ, SL7NZ and SL72P can support EM64T. The product with suffix P supports hardware antivirus, EM64T and enhanced temperature control function, and has 2MB secondary cache. Careful readers will find that this is Intel's new 6XX series processor. Personally, I think that as long as it is an E0 core processor, it should have these three functions except the size of the second-level cache. Just to distinguish between 6xx and 5xx series, Intel artificially controlled P4J without EM64T, and enhanced the temperature control function. Here, the author once again emphasizes the choice of E0 stepping Prescott, and it is very likely to open these functions by upgrading the BIOS in the future.
● Chaotic Prescott processor
Although Northwood processor has the difference of A/B/C, it is easy to identify. Although there are several processor steps such as B0/C 1/D 1/M0, D 1 is usually sold in the market, so it will not be confused as long as it is slightly differentiated according to the above method when purchasing. However, Prescott is rather chaotic. Only Prescott with "1M/800" has multiple E/J/F/P suffixes. Besides whether you are familiar with supporting hyper-threading, EM64T and hardware antivirus, there are several differences that you don't understand.
First of all, the difference of power specifications: FMB 1.5 and FMB 1.0 (socket 478 Prescott only), which also needs to be understood through S-Spec; Followed by the maximum power consumption: 04A and 04B (LGA 775 only). In the latest LGA 755 product line, Intel has formulated two power consumption schemes. 4A is the mainstream scheme with less power consumption and slightly poor performance. 04B is called a high-performance scheme with high power consumption and strong performance. Intel directly indicates whether it is 04A or 04B on the processor box, so it is easy to distinguish. Of course, it is more accurate to use S-Spec to distinguish.
To sum up, the difference between Prescott processors must be based on S-Spec. At the end of this paper, the S-Spec number of the known Prescott processor is listed in detail for your reference.
● Pay attention to the tread of Celeron D..
Celeron D includes C0/D0/E0, and D0 Celeron D 3 15 or 320 is the most popular in the market. The new E0 stepping LGA 775 Celeron D is called Celeron D J, which supports hardware antivirus. Choosing Celeron d still needs to refer to S-Spec. For example, Celeron D 3 15 is the lowest frequency multiplier in this series, and it has strong overclocking ability. It also includes a variety of stepping products, such as SL7XG is C0 stepping, SL7XY/SL7WS is D0 stepping, SL8AW/SL87K is E0 stepping, and E0 stepping is the first choice. Please refer to the list at the end of the article for other models.
So far, the author has comprehensively analyzed the serial numbers of various Intel and AMD processors available in the current market (including the second-hand market), and figuring out the difference between these serial numbers means that you will become an expert in processor identification. In addition, another great significance of fully understanding the processor number is to find the processor that is easier to overclock through the step value. Below, the author lists the number of common processors in the market, the number of Intel product list S-Spec and AMD product list OPN.
Table 1: Intel Northwood s specifications
The author chooses an S-Spec with different steps at each main frequency as a reference. Others can be found in the query or downloaded from Intel's official documentation area/design/pentium4/documentation.htm.
Processor name S-Spec stepping core voltage
1.6 GHz P4A sl 668 b 0 1.5
1.8 GHz P4A SL63X B0 1.5
SL6QL c 1 1.475 ~ 1.525
SL6PQ D 1 multi-voltage
2.0GHz P4A SL5YR B0 1.5
SL6E7 C 1 1.525
SL6PK D 1 Multi-voltage
2.2GHz P4A SL5YS B0 1.5
SL6E8 C 1 1.525
SL6QN D 1 multi-voltage
2.26GHz P4B SL67Y B0 1.5
SL6RY C 1 1.53
Sl6pbb d11.525 (multi-voltage)
2.4GHz P4A SL65R B0 1.5
SL6S9 C 1 Multi-voltage
SL6QP D 1 multi-voltage
2.4GHz P4B SL67Z B0 1.5
SL6RZ C 1 1.53 (Multi-voltage)
SL6PC D 1 1.525 (multi-voltage)
2.4GHz P4C SL6WR D 1 multi-voltage
2.5 GHz P4A SL6EB c 1 1.525
SL6QQ D 1 multi-voltage
2.53GHz P4B SL682 B0 1.5
SL6DW C 1 1.525
SL6PD D 1 1.525 (multi-voltage)
2.6 GHz P4A SL6GU c 1 1.5
SL6QR D 1 Multi-voltage
2.6GHz P4C SL6WH D 1 multi-voltage
2.66 GHz P4B SL6DX c 1 1.525
SL6QA D 1 1.53 (Multi-voltage)
2.8 GHz P4A SL7EY d 1 1.475 ~ 1.55
2.8 GHz P4B SL6HL c 1 1.525
SL6K6 C 1 1.525
SL6QB D 1 1.53 (Multi-voltage)
2.8GHz P4C SL6WJ D 1 multi-voltage
3.0ghz p4csl6KWD1multi-voltage
3.06 GHz P4B SL6JJ c 1 1.525
SL6PG D 1 1.55 (multi-voltage)
3.2 GHz P4C SL6WE d 1 1.25 ~ 1.4
3.4 GHz P4C SL7AJ C0( 1MB L2) 1.25 ~ 1.4
sl 793d 1 1.25 ~ 1.4
Intel Celeron d
Table 2: Intel Celeron D S specifications
Celeron D has few models, but it has steps of C0, D0 and E0. At present, there are still many C0 stepping products in the domestic market, especially boxed products. If overclocking is required, it is recommended to select products in batch D0 or E0.
Processor name S-Spec stepping interface
Celeron D 3 15 SL7XG C0 socket 478
SL7WS D0 socket 478
SL8AW E0 socket 478
Celeron D 320 SL7C4 C0 socket 478
SL7JV D0 socket 478
SL87J E0 socket 478
SL7VQ E0 LGA 775
Celeron D 325 SL7C5 C0 socket 478
Sl7sd0 socket 478
SL7NU E0 socket 478
SL7VR E0 LGA 775
Celeron D 330 SL7C6 C0 socket 478
SL7ST D0 socket 478
SL7NV E0 socket 478
SL7VS E0 LGA 775
Celeron D 335 SL7C7 C0 socket 478
SL7Q9 D0 socket 478
SL7NW E0 socket 478
SL7VT E0 LGA 775
Celeron D 340 SL7Q9 D0 socket 478
SL7TS E0 socket 478
SL7VV E0 LGA 775
Celeron d345sl7dd0 socket 478
SLYW3 E0 socket 478
SL7TQ E0 LGA 775
Intel Prescott article
Table 3: Intel prescaler specifications
Prescott's situation is complicated, so the author tries to list the known S-SPECs. It should be noted that Socket 478 product (indicated by S in the table) has no power consumption, while LGA 775 product (indicated by L in the table) has no power specification.
Processor Name S-Spec Stepping EM64T Power Specification Hyper-Threading Power Interface
2.26 GHz P4A SL7D7 (512kl2) C0No. FMB 1.0 is not applicable.
2.4GHz P4A SL7E8 C0 is not applicable without FMB 1.0.
SL7YP D0 is not applicable without FMB 1.0.
2.4GHz P4E SL7FY C0 is not applicable without FMB 1.0.
2.66ghz p4a sl7pte0No N/A No 04A L
2.8GHz P4A SL7D8 C0 is not applicable without FMB 1.0.
SL7E2 D0No. FMB 1.0 is not applicable.
SL7K9 D0 No FMB 1.0 Unknown Not applicable.
SL7PK E0No. FMB 1.0 is not applicable.
SL7J4 D0 is either not applicable or not applicable.
SL7KH D0 No N/A unknown 04A L
2.8GHz P4E SL79K C0No. FMB 1.0 is not applicable.
SL7E3 D0 is not applicable without FMB 1.0.
SL7KA D0 is not applicable without FMB 1.0.
FMB 1.0 not applicable.
SL7J5 D0 without N/A is 04A L.
SL7KJ D0 None, not applicable to 04A L.
SL82V E0 is not applicable to is 04A L.
SL7PR E0 None, not applicable to 04A L.
2.93 GHz P4A SL85E0No. 04A L not applicable.
3.0GHz P4E SL79L C0 is not applicable without FMB 1.0.
SL7L4 D0 without FMB 1.0 is not applicable.
SL7E4 D0 is not applicable without FMB 1.0.
SL7KB D0 without FMB 1.0 is not applicable.
SL7PM E0No. FMB 1.0 is not applicable.
SL7J6 D0 without N/A is 04A L.
SL7KK D0 No N/A is 04A L
SL82X E0 is not applicable to is 04A L.
SL7PU E0 None, not applicable to 04A L.
3.06 GHz P4A SL87LEE0No. 04A L N/A
3.2GHz P4E SL7B8 C0 is not applicable without FMB 1.5.
SL7L5 D0 without FMB 1.0 is not applicable.
SL7E5 D0 is not applicable without FMB 1.0.
SL7KC D0 FMB 1.0 is not applicable.
SL7J7 D0 without N/A is 04A L.
SL7KL D0 is not applicable to 04A L.
SL7LA D0 is N/A is 04A L.
FMB 1.0 not applicable.
SL7PW E0 None, not applicable to 04A L.
Sl7pe0 is N/A is 04A L.
SL82Z E0 None, not applicable to 04A L.
3.4GHz P4E SL7B9 C0 is not applicable without FMB 1.0.
SL7E6 D0 is not applicable without FMB 1.5.
SL7KM D0, not applicable to 04B L.
SL7L8 D0 is N/A is 04B L.
SL7J8 D0 without N/A is 04B L.
SL7PP E0No. FMB 1.0 is not applicable.
FMB 1.5 not applicable.
Sl7pie0 No, 04A L is not applicable.
SL7PZ E0 not applicable, 04A L.
SL833 E0 None, not applicable to 04A L.
3.6 GHz P4E SL7J9D0No N/A is 04B L
Sl7kND0None, 04B L is not applicable.
SL7L9 D0 is N/A is 04B L.
SL84E0 is not applicable, not applicable to 04B L.
SL7Q2 E0 No, not applicable to 04B L.
SL7NZ E0 is N/A is 04B L.
3.8 GHz P4E SL82E0No N/A is 04B L
SL84Y E0 None, not applicable to 04B L.
SL72P E0 is N/A is 04B L.
Each processor has a number, which can reflect the main frequency, front-end bus frequency, secondary cache, working voltage and other parameters. Reading this number can not only identify the processor, but also prevent fakes to some extent when actually buying.
First, display the processor number.
1. Intel processor
At present, the Intel processors on the market mainly include Pentium 4 and Celeron D series, and the old Celeron based on Northwood Core is tending to be eliminated. The surfaces of these processors are covered with metal heat sinks, and the processor numbers are on them.
The serial number of the Pentium 4 processor surface.
Note: The logos of all Intel processors are similar. Even if there are occasional adjustments, the arrangement order is only fine-tuned, but the basic information remains unchanged.
As can be seen from the above figure, the first line is marked as the basic parameters of the processor, which is expressed in the form of "main frequency/secondary cache/front-end bus frequency/voltage (some voltages are not marked)" (this article is called "simple number"). This line of information is especially useful for beginners to understand the basic parameters of the processor.
The second line is S-Spec and its origin. S-Spec contains more secrets of Intel processors. This five-digit number can fully understand the main frequency, secondary cache, FSB frequency, core voltage, temperature, processor step value and other information. Although you can't directly see the meaning of S-Spec, it is the most useful tool for choosing Intel processors. The author will introduce it in detail later, and list the common Intel processor S-Spec at the end of the article for reference. After S-Spec comes the origin of processors, which are common in Malaysia, Costa Rica and China.
The third line is FPO and serial number, which is the unique factory number of each processor. Consumers who buy boxed processors should pay attention to whether the FPO number on the outer packaging is consistent with the processor, and can confirm whether it is a real boxed product through Intel 800 phone.
1958, kilby, an engineer from Texas Instruments, made the world's first integrated circuit by integrating discrete components such as resistors and capacitors on a semiconductor silicon wafer. Because of this, the 2000 Nobel Prize in Physics was awarded to retired kilby. 1959, Noyce of American Fairchild Company made semiconductor integrated circuits by plane technology, which opened an era when integrated circuits were more attractive than gold. Later, Moore, Noyce and Grove left Fairchild and started their own business together. The three agreed that the most promising semiconductor market is the computer memory chip market. Another important reason to attract them to set up a new company is that this market is almost entirely dependent on high technology. You can put as many circuits as possible on one chip, and whoever has high integration will become the leader in this industry. Based on the above considerations, Moore named the new company Intel, which is a combination of two English words "Integrated Electronics", symbolizing that the new company will prosper in the integrated circuit market, and the result is really like this. It seems that in Moore's lifetime, it will develop to let him give a name.
At that time, the three entrepreneurs persuaded venture capitalist Arthur Rock to invest $2 million in them. They also found the best place to start a business, that is, the building of the former Union Carbide Electronics Company, which is much better than HP's garage. Shortly after the company was founded, the three founders and company employees (at this time, it was the end of 1968, and Intel had agreed not to stick to any specific technology or product line. In Noyce's words, they will "take a quick shot of all technologies today, find out which technology is effective, which technology is the most effective, and develop which technology". The company has enough time, talents and funds, so they can't make a move. Noyce said: "There is no contract that we must guarantee the production of a certain production line. We are not bound by any old technology. "
Intel found that when electrons appear or disappear in a tiny part of an integrated circuit block, many bits of information (the smallest unit of measurement of data) can be stored on a silicon chip of a micro-integrated circuit very cheaply. They first applied this discovery to business. 1969 In the spring, after the first anniversary of the company's establishment, Intel Corporation produced the first batch of products, namely bipolar processing 64-bit memory chips. Soon, the company introduced a 256-bit MOS memory chip. A small Intel company, with its two new products, broke into the whole computer memory market-a brilliant start, while other companies could not produce MOS chips and bipolar chips until 1980.
As Japanese companies join the competition, the storage business is becoming more and more difficult. Although many Americans complained at that time that Japanese companies dumped products to the United States at a price lower than the cost, there is no denying the fact that Japan's speed and quality in chip manufacturing are unparalleled. At this time, Intel is facing the biggest survival crisis in history. But in the end, they made an admirable decision: give up memory and join the microprocessor business.
Speaking of microprocessor business, it was actually a very accidental thing at first: a customer of Intel (Busicom, a Japanese manufacturer that no longer exists) asked Intel to design some processing chips specially for it. During the research, Intel researcher Hoff asked himself: Can integrated circuits be operated by external software with simple instructions? Why can't all the logic on this computer be integrated into a chip and a simple and general program can be compiled on it? This is actually the principle of all microprocessors today. But Japanese companies have no interest in this. With the help of colleagues and the support of the company, Hoff integrated all the functions of the central processing unit on one chip, plus memory; This chip, later called 4004, was perfected, which is the first microprocessor in the world.
197 1 year, Intel gave birth to the first microprocessor-4004. The chip is actually specially designed and manufactured for Busicom calculator, but you can already see the shadow of personal computer in it. It is said that at that time, an American with long hair saw the news of I4004 in a radio magazine and immediately wanted to develop a personal operating system with this CPU. As a result, after some careful tossing, I found that the function of I4004 was too weak, and the system function and Basic language he wanted to realize could not be realized on it, so he had to give up. This man is Bill Gates, the boss of Microsoft. But since then, he has been very concerned about Intel's trends, and finally achieved 1975 Microsoft.
Next, 8008, 8008 has twice the computing power of 4004. 1974, a radio magazine published a machine using 8008 as the processor, named "Mark-8", which is also the earliest known home computer. Although it is very difficult to use, control, program and maintain Mark -8 from today's point of view, it was a great invention at that time.
The next generation of products is called 8080, 8080, which is used in computers with Altair (Altair, whose name comes from a popular science fiction drama on TV at that time). This is also the first well-known personal computer in history. At that time, the price of this computer suite was $395, and the sales performance reached tens of thousands in just a few months, creating a milestone in the history of personal computer sales.
The integration of 4004 has only 2300 transistors, and its function is actually very weak and its operation speed is slow, so that it can only be used in Busicom calculator, not to mention complicated mathematical calculation. However, compared with the first electronic computer ENIAC, it is much lighter. And the greatest historical significance is that it is the first general-purpose processor, which is a rare breakthrough in the era when ASIC design is rampant. The so-called ASIC design is to design unique products for different applications. Once the application conditions change, it needs to be redesigned. Of course, in terms of commercial profit, it is very beneficial to design companies. However, Intel's vision is not so short-sighted. Hoff put forward a bold idea: using general hardware design and external software support to complete different applications, which is the original idea of general microprocessor.
Intel quickly proved the idea and found it feasible. Moreover, the advantage of this product is that it can complete different tasks with the support of different software, which is much simpler than redesigning ASIC. Seeing the broad prospect of this product in the future, Intel immediately put into the design work, and soon launched this product-the world's first microprocessor Intel 4004.
In fact, 4004 processing can only process 4 bits of data, but the internal instruction is 8 bits. 4004 has 46 instructions and is packaged in 16 pin in-line package. Data memory and program memory are separated, 1K data memory and 4K program memory. The running clock frequency is expected to be 1M, and it will eventually reach 740kHz, which can perform binary coded decimal mathematical operations. This processor was quickly recognized by the whole industry, and the blue giant IBM also equipped the 4004 on the IBM 1620 machine.
Shortly after the release of 4004, Intel released several CPU:4040 and 8008. The market response was mediocre, but it laid a good foundation for the development of 8-bit microprocessors. 1974, Intel developed an 8080 processor based on 8008, which has 16-bit address bus and 8-bit data bus, including seven 8-bit registers (A, B, C, D, E, F, G, among which BC, DE and HL can be combined to form1.
1978,8086 processor was born. This processor marks the beginning of the x86 dynasty. Why celebrate the 25th anniversary of Intel x86 architecture? The main reason is that since 8086, the most widely used PC industry foundation has been established. Although it has been 32 years since 197 1, Intel has made 4004. But there has never been such a far-reaching masterpiece as 8086.
Another more critical factor is that IBM is studying new PCs to attack Apple's personal computers. IBM needs to choose a powerful and scalable processor to drive it. Intel's x86 processor won an absolute victory and became the new "brain" of IBM PC. This historical choice also made Intel among the top 500 companies in the world in the future, and was called "one of the 70 business miracles" by Fortune magazine.
The success of IBM PC not only brought prosperity to Intel's business, but also created another business miracle-Microsoft. Bill Gates hitchhiked to sell the DOS operating system, which earned him the first pot of gold for dominating the software industry today. Not only that, third-party vendors such as Compaq (now a part of HP) also benefited greatly from IBM's foresight and the opening of PC architecture license. Even the economic take-off of Taiwan Province Province has an inevitable connection with this historical combination. This event is very commendable from both historical and industrial perspectives!
In fact, IBM chose the 8088 model in PC XT. Technically, 8088 is actually a simplified version of 8086. Its internal instruction is 16 bits, but its external instruction is 8-bit data bus. Compared with the specifications of 8086 internal data bus (CPU internal data transmission bus) and external data bus (CPU external data transmission bus), both of them are 16 bits, and the address bus is 20 bits, which can address the memory of 1MB, which is a little worse, but it is enough for DOS systems and applications at that time. 8086 integrates 29,000 transistors with a clock frequency of 4.77MHz, and also produces a mathematical coprocessor 8087. These two chips use the same instruction set and can cooperate with each other to improve the efficiency of scientific operation.
Of course, the current CPU has built-in mathematical coprocessor, so there is no need for additional mathematical coprocessor chips. However, due to the technical limitations in the 1970s, the mathematical coprocessor can only be made into another chip for users to choose from. This helps to reduce the manufacturing cost, increase the output and reduce the expenses of users who are insensitive to speed: they can temporarily not buy a math coprocessor until they need to buy one and plug it into an IC socket.
1982, Intel released the 80286 processor, which is 286. This is the first Intel processor that can run all processors. In the six years since its release, 286-based15 million PCs have been delivered worldwide.
The 80286 chip integrates 143000 transistors with a word length of 16 bits, and the clock frequency is gradually increased from the initial 6MHz to 20MHz. Its internal and external data buses are 16 bits, and its address bus is 24 bits. Compared with 8086, the addressing ability of 80286 is 16MB, and a large amount of storage space can be simulated by external storage devices, thus greatly expanding the working range of 80286. It can also make the processor quickly switch back and forth between various tasks through the multi-task hardware mechanism, so as to run multiple tasks at the same time, which is five times or more faster than 8086. The 80286 in AT machine used by IBM has more advanced technology. Compared with IBM PC, the external bus of AT machine is 16 bits (8 bits for PC XT machine), and the memory can generally be expanded to 16MB, which can support larger hard disk and VGA display system, and has made great progress in performance compared with PC XT machine.
However, at this time, there was a great disagreement within IBM: many people inside opposed the rapid transition to 286 computer sales, because 286 PC would have an impact on IBM minicomputers and previous PC XT sales, and they hoped for a slow transition. But intel can't wait, the 80286 processor has been mass-produced, and it is impossible to pile it in the warehouse and wait for IBM to digest it slowly; At this time, Compaq, which produces IBM-compatible PCs, exploited a loophole-quickly launched 286 PCs and defeated IBM to become the new overlord of the PC market.
Microprocessor determines the performance and speed of computer. Whoever can make a high-speed PC with excellent performance will lead the new computer trend. This is the rule of the game. IBM people initially followed this rule, so they achieved great success in the PC market, but in the 286 era, they gave up the right choice, which is really a pity.
80386 entered the 32-bit generation.
1985, Intel introduced the 80386 processor again. 386 integrates 275,000 transistors, which is 100 times more than 4004 chips. 386 is also the first 32-bit processor of Intel and the first processor with multi-tasking function, which has an important impact on the development of Microsoft operating system. The so-called "multi-tasking" means that the processor can process instructions of several programs at the same time.
However, just like the transition to 286, Intel has also encountered great pressure. At that time, there was a popular view that 286 was enough, there was no need to produce 386 computers, and the sales volume was not ideal at first. However, Intel's leaders do not think so. It is refreshing to adopt many new methods in propaganda and learn from many consumer products. On the other hand, 386 chips are divided into different specifications to meet the needs of different users. Especially for the 80386SX chip, the internal data bus is 32 bits, which is the same as 80386, but the external data bus is 16 bits, which has the advantage of 386 and the cost advantage of 286, and has achieved great market success. At the same time, the original 386 chip was renamed as 386DX to distinguish it from 386SX.
In the 386 era, Intel made great progress in technology. 80386 has 275000 transistors, and the clock frequency is 12.5MHz, and then it is increased to 20MHz, 25MHz and 33MHz. The internal and external data buses of 80386DX are 32-bit, and the address bus is also 32-bit, which can address up to 4GB of memory. In addition to real mode and protected mode, it also adds a working mode called virtual mode, which can simulate multiple 8086 processors at the same time and provide multi-task capability.
1989, Intel released the 486 processor. The 486 processor is a very successful commercial project of Intel. Many manufacturers have also seen the development law of Intel processors, so they quickly transformed successfully with the help of Intel's marketing war. The 80486 processor integrated 1.25 million transistors, and the clock frequency gradually increased from 25MHz to 33MHz, 40MHz, 50MHz and later 100Mhz.
80486 is also the first CPU with digital coprocessor in Intel, and RISC (Reduced Instruction Set) technology is used for the first time in x86 series, thus improving the speed of executing instructions in each clock cycle. 486 also adopts the burst bus mode, which greatly improves the data exchange speed between the processor and the memory. Because of these improvements, the performance of 80486 is more than four times faster than that of 80386 with 80387 math coprocessor.
Intel once again applied the strategy of cell users to 486 products, so 486 was divided into 486DX with mathematical coprocessor and 486SX without mathematical coprocessor, and the price of 486SX was cheaper. Later, the specifications of 486 were improved in frequency doubling, and new "variants" of 486DX2 and 486DX4 appeared. Take DX2 as an example, that is, in the processor.