Processors
A Central Processing Unit (CPU), or processor, is the "brain" of a computer. The processor is the component that handles all mathematical operations, and all data transfers. Without a processor, there could not be a computer. The processor has been improved in terms of size and speed throughout its existence. In the early '50s, a computer (of which the processor was a large part) would take up an entire large room. If it could run software similar to Microsoft Windows on it, it would take a couple months to load. Those processors would only contain a couple thousand vacuum tubes (a vacuum tube has the same in function as transistors). Comparing that to one of Intel's later models shows some significant improvement. The chip size has been reduced to less than a square inch, and an entire computer can fit into a relatively small box. To load up Microsoft Windows takes up between 30 seconds and 1 minute, and the chip contains about 7 million transistors.
Since it may be of interest to many of you, we have presented the history of development of several common CPU lines. Unfortunately, we were only able to find ample information about the x86 line, but not the other lines.
The first x86 processor was developed in 1979 by Intel and was called the 8088. This processor then went through six different generations. Tens of millions of these x86 processors are now being sold each year. Numerous clones of these x86 processors have been made but only Cyrix, AMD, and Centaur are still able to compete.
The 8088 and the 8086 could run the same programs but one couldn't fit into the other's socket. The 8086 got its success mainly from the IBM-PC although IBM preferred implementing the 8088 into its computers because it was less expensive and less complex. When the IBM-PC appeared, Intel Corporation was only at its beginnings therefore, it couldn't produce it's processors quickly enough for an expanding company such as IBM. Therefore, to increase the production of processors, Intel sold the fabrication rights to several companies including AMD, Harris, Hitachi, IBM, and Siemens. Depending on the manufacturer, the 8086 and 8088 processors would run at a speed ranging from 4MHz to 16MHz.
The 8086 developed its clones. The first were made by NEC in 1985. The 8088's clone was the V20, the 8086's clone was the V30. These ran roughly 20% faster than the 8088 and 8086 when working at the same clock speed (i.e.MHz).
The 80186 and the 80188, were very advanced for their time but they were not adopted by hardware manufacturers. Though this chip didn't have much success as the main CPU it enjoyed a huge success as an embedded processor (a processor which you cannot remove, usually found in small electronic components such as a disk driver or disk controller). You might still find some of these processors in your computer. Depending on the manufacturer, you could find 80186/80188 processors running at a speed ranging between 6 and 40MHz.
This was the first time that Intel didn't produce simultaneously a 16-bit bus version (ending with a 6) and an 8-bit bus version (ending with an 8). The 80286's speed could range between 6MHz and 25MHz depending on the manufacturer. It didn't go faster than the 80186 but it had several improvements (a little too complex to explain). This processor was used in the IBM PC-AT in 1985 (three years after the introduction of the 80286).\
The 80386 was introduced by Intel in 1985. It had several significant improvements compared to its ancestors (it was a 32-bit processor). Although it wasn't adopted immediately by computer manufacturers, it enjoyed a long life span. The 80386 was later renamed 80386 DX, for Intel developed another 80386 processor, the 80386 SX. This was a downgraded version of the 80386 DX. Intel developed other processors belonging to the 80386 family, the 80376 and the 80386 LS. The success of the 80386 led other companies to emerge on the processor market, developing their own 80386 clones, such as the Am386, the 486 by Cyrix (this was a 386 pin-compatible CPU but was called a 486, a nomenclature Cyrix still uses). This competition lead Intel to use some business practices which many considered as anti-competitive or even illegal.
The 80486's major improvement over the 80386 was the integration of an 80387 math coprocessor into the 80486. Intel followed the pattern of the 80386 for the 80486 by introducing a 80486SX and renaming its original processor 80486DX. The difference between the two was the presence or absence of the math coprocessor (in the earlier versions of the SX, the math-coprocessor was present but didn't function). Intel then introduced the 80487SX (a math coprocessor) which was used as an expensive upgrade for the 80486SX making it a 80486DX. This 80487SX was in fact a fully operational 80486DX on which a few pins had been relocated so as not to allow the customer to use the cheaper 80486DX as an upgrade. The 80487SX in fact electronically disabled the 80486SX and just took its place making the 80486SX a useless space consumer. Unfortunately the customer never realized or suspected the existence of this twisted game established by Intel. Several other versions of the 80486 were introduced, the 80486SL, the 80486DX2 and the 80486DX4 (these doubled and tripled the core clock frequency). The 80486 also had its clones such as the AMD Am486 and the Cyrix 5x86.
The Pentium was a big step in the evolution of the processor. The major difference was that it contained more than one execution unit. An execution unit receives the instructions (see the Instructions and Opcodes page) and activates them. Having more than one execution unit meant that the processor could do more than one thing at a time. Intel had by now expanded enough to provide the market with a sufficient amount of processors and therefore it didn't want to reveal the fabrication secrets of the Pentium.
The Pentium came in different versions: 60,66,75,90,100,120,133,166,200, and 233 MHz. Then Intel introduced the MMX-enhanced processors (see MMX). AMD developed the K5 processor around this time but this one was a deception. After this failure, AMD bought Nexgen which had already developed its own x86 compatible processor, the Nx586. Nexgen also had just finished designing the basis of their next generation processor the Nx686. This was used as a working base for the very succesfull AMD K6 processor. At this time, Cyrix developed its 6x86 (pin-compatible with the Pentium) which then received MMX technology. Centaur Technology made its appearance on the processor market at this time with the sub-$1000PC (a relatively low-price Pentium compatible chip) which was received with a great success.
The Pentium Pro or P6 didn't benefit from major technical improvements. Intel didn't try to keep its fabrication a secret since by now the Pentium's secret programming features had been revealed. The performance of the Pentium Pro was nicely improved by a large L2 cache which had the disadvantage of being expensive. Intel's attempts to eliminate further competition took another step with the Pentium Pro. Intel gained patent protection for some of the Pentium Pro's pins making it very difficult to clone legally. Intel also claimed that the Pentium Pro's 2nd-level cache could not work at more than 200MHz and used this as a pretext for the launching of the Pentium II. AMD and Cyrix stayed with a Pentium compatible pin-out continuing to develop their respective processors.
For the Pentium II, Intel totally abandoned sockets and turned to the concept of slots. Intel introduced the slot-1 whose secrets were hidden by patent protection. Slot-1 was supposedly introduced for technical reasons. However these reasons disappeared with the decline of the Pentium Pro. Intel then claimed that they had managed the "breakthrough" that they had qualified earlier on as being impossible. They then announced the introduction of a Pentium II running at the speed of 400MHZ with a full-speed 2nd-level cache (2nd-level cache couldn't run faster than 200MHz according to them).
The Celeron was created as an attempt to regain some market share lost due to the success of the sub-$1000PC. The Celeron was in fact a Pentium II whose 2nd-level cache has been removed making it supposedly twice as slow as a Pentium MMX. When Intel realized this, they put in 128K L2 cache, making the processor faster.
The Xeon is a Pentium II whose 2nd-level cache runs at full processor speed (a regular Pentium II's 2nd-level cache runs at half the processor speed). Xeon is not slot compatible with the Pentium II but uses a new type of slot named Slot-2.
These will be presented as a timeline, as less information was available.
- Shipping of the 50-MHz Sun SuperSPARC processors.1994: The 64-bit UltraSPARC RISC processor is revealed.
- Shipping of the 60-MHz Sun SuperSPARC processor.
- Shipping of the 85-MHz and 110-MHz MicroSPARC II processor.
-Shipping of the 85-MHz Sun SuperSPARC II processor.
-Sample availability of the 64-bit UltraSPARC processor announced.
-Shipping of the 143-MHz and 167-MHz UltraSPARC processor.
Motorola processors were mainly used in Apples/Macintoshs.
1974: Introduction of Motorola's early 8-bit microprocessor, the 6800 chip. It was used in microcomputers and industrial and automotive control devices. It was designed by Chuck Peddle and Charlie Melear.-Appearance of the 68020, a 32-bit processor.1985: The 68008 CPU chip is revealed
-The 68010 CPU chip is revealed.
-Development of the 88000 processor begins.1987: The 68030 is revealed.
-The 68030 is announced by Motorola.
-The 68040 is announced by Motorola.1990:
-A 50-MHz version of the 68030 is announced.
-The 32-bit 25-MHz 68040's availability is announced.1991: The 68040 is available ( after a year's delay ).
-A new line of single-chip RISC processors is announced by Motorola, the first of which is to be the 88110.
-The first PowerPC 601s are shipped.1994:
-The second chip in the PowerPC line, the 603, is starting to be produced. It is also introduced in a 66-MHz and 80-MHz version.
-An 80-MHz version of the PowerPC 601 is introduced.
-The 100-MHz PowerPC 601 is announced.1995: The MC68328 chip is introduced by Motorola.
-The 100-MHz PowerPC 604 is announced.
-The 133-MHz PowerPC 620 is announced and a prototype is introduced.
-Production of the PowerPC 604 starts.
-A 120-MHz PowerPC 601 is shown running in a prototype Power Macintosh.
-The 8-MHz R2000 32-bit CPU is revealed.1991:
-Volume shipments of the R2000 processor are started.
-The R4000 RISC processor architecture is revealed.1992: The 100-MHz R4000 processor is shipped.
-Samples of the R4000 processor are starting to be shipped.
-The 100-MHz R4000 ( 64-bit RISC processor ) is officially introduced.
-The availability of the 150-MHz 64-bit R4400 RISC is announced.1994:
-The 100-MHz R4600 RISC microprocessor is shipped.
-Availability of the 200-MHz 64-bit R4400 RISC microprocessor.1995:
-The 75-MHz R8000 RISC microprocessor is shipped.
-The 133-MHz R4600 RISC microprocessor is shipped.
-The 200-MHz 64-bit R4400 RISC microprocessor is shipped.
-The R10000 RISC microprocessor is announced.
-The 90-MHz R8000 RISC microprocessor is shipped.1996:
-The 250-MHz R4400 RISC microprocessor is shipped.
-The 200-MHz R5000 RISC microprocessor is shipped.[an error occurred while processing this directive]
-The 200-MHz R10000 RISC microprocessor is shipped.