Development path of personal computer processor

' The path ahead of the desktop processor (processor) , somewhat shows us the future PC processor. Computer processors have changed a lot from the beginning, from Intel's 4bit 4004 processor, AMD's AM2900 in the early 1970s to today's 64-bit processors.

From Intel's first PC (PC) processor to 4-core, 8-core PC processors in 2007 and more in the future. We will take a look back at the 36-year-old development of the processor that is full of stiff competition, and look at the technologies applied in the PC processor.

4bit processor

The 4004 was the first processor introduced by Intel in November 1971, used in Busicom's calculator. The 4004 has a speed of 740KHz, capable of handling 0.06 million instructions per second (MIPS instructions per second - MIPS); produced on 10 µm technology, has 2,300 transistors (transistors), memory extends to 640 bytes.

4040 , an improved version of the 4004 was introduced in 1974, with 3,000 transistors, speeds from 500 KHz to 740KHz.

8bit processor

8008 (1972) used in Terminal Datapoint 2200 of Computer Terminal Corporation (CTC). 8008 has a speed of 200kHz, manufactured on 10 µm technology, with 3,500 transistors, memory expandable to 16KB.

8080 (1974) used in the Altair 8800 computer, which has 10 times the speed of 8008 (2MHz), manufactured on 6 µm technology, capable of handling 0.64 MIPS with 6,000 transistors, with 8 data bus bits and 16 address bus bits, memory expandable to 64KB.

8085 (1976) used in scale Toledo and peripheral control devices. The 8085 has a speed of 2MHz, manufactured on 3 µm technology, with 6,500 transistors, has 8 data bus bits and 16 addressable bus bits, 64KB of expandable memory.

16bit processor

8086 appeared in June 1978, used in mobile computing devices. 8086 is produced on 3 µm technology, with 29,000 transistors, has 16 data bus bits and 20 address bus bits, 1MB of expandable memory. Versions of 8086 include 5, 8 and 10 MHz.

8088 was released in June 1979, and was chosen by IBM for its first PC (PC); This also helps Intel become the largest computer processor manufacturer in the world. 8088 is identical to 8086 but has command line management capability. 8088 also uses 3 µm technology, 29,000 transistors, 16-bit internal architecture and 8 external data bus bits, 20 address bus bits, memory up to 1MB. Versions of 8088 include 5 MHz and 8 MHz.

80186 (1982) is also called iAPX 186. Mainly used in embedded applications, terminal controllers. Versions of 80186 include 10 and 12 MHz.

80286 (1982) is known as 286, the first Intel processor to run all applications for previous processors, used in IBM PCs and compatible PCs. 286 has 2 operating modes: real mode with DOS program in 8086 simulation mode and cannot use more than 1 MB RAM; Protect mode increases processor performance, can access up to 16 MB of memory.

286 uses 1.5 µm technology, 134,000 transistors, expandable memory to 16 MB. Versions of 286 include 6, 8, 10, 12.5, 16, 20 and 25MHz.

32bit processor

Intel386 includes 386DX, 386SX and 386SL families. Intel386DX is the first 32-bit processor Intel introduced in 1985, used in IBM PCs and compatible PCs. The Intel386 is a big leap forward from previous processors. This is a 32-bit processor capable of multitasking, it can run many different programs at the same time. The 386 uses 32-bit registers, which can transmit 32 bits of data at the same time on the data bus and use 32-bit data to determine the address. Like the 80286 processor, the 80386 works in two modes: real mode and protect mode.

386DX uses 1.5 µm technology, 275,000 transistors, memory expandable to 4GB. The 386DX versions include 16, 20, 25 and 33 MHz (1 µm technology).

386SX (1988) uses 1.5 µm technology, 275,000 transistors, 32 bit internal architecture, 16 external data bus bits, 24 address bus bits, 16MB expansion memory; including versions 16, 20, 25 and 33 MHz.

386SL (year 1990) was designed for mobile devices, using 1 µm technology, 855,000 transistors, 4GB expandable memory; including 16, 20, 25 MHz versions.

486DX was born in 1989 with a 32-bit data bus structure. 486DX has 8 KB of primary memory (L1 cache) to reduce data waiting time from memory, built-in math co-processor. In addition, 486DX is designed to be a pipeline, which can handle an instruction in a clock.

486DX uses 1 µm technology, 1.2 million transistors, 4GB expandable memory; including 25 MHz, 35 MHz and 50 MHz (0.8 µm) versions.

486SX (1991) used in low-end computers, with the same design as the 486DX system but not integrated with the co-processor. 486DX uses 1 µm technology (1.2 million transistors) and 0.8 µm (0.9 million transistors), 4GB expandable memory; including versions 16, 20, 25, 33 MHz.

486SL (1992) was the first processor for laptops (laptops), using 0.8 µm technology, 1.4 million transistors, 4GB expandable memory; including versions 20, 25 and 33 MHz.

Intel Pentium, the next generation 486 processor was born in 1993. The Pentium's biggest innovation is to design two lines of pipeline, internal data is capable of executing two instructions in a cycle, so the Pentium can handle instructions twice as much as 80486 DX at the same time. The 16KB primary memory contains 8 KB of data and another 8 KB to contain the command. Improved math co-processor helps increase computing power for applications.

Pentium using 0.8 µm technology contains 3.1 million transistors, with speeds of 60, 66 MHz (socket 4 273 pins, PGA). Versions 75, 90, 100, 120 MHz using 0.6 µm technology contain 3.3 million transistors (socket 7, PGA). Versions 133, 150, 166, 200 use 0.35 µm technology containing 3.3 million transistors (socket 7, PGA)

Note : Readers can consult more information about processor sockets at http://en.wikipedia.org/wiki/List_of_CPU_sockets.

Pentium MMX (1996), an improved version of Pentium with MMX technology developed by Intel to meet the needs of multimedia applications and communications. MMX combined with SIMD (Single Instruction Multiple Data) allows processing multiple data in the same instruction, increasing processing capability in graphics and multimedia tasks.

Pentium MMX uses 0.35 µm technology containing 4.5 million transistors, with speeds of 166, 200, 233 MHz (Socket 7, PGA).

Pentium Pro . Following the success of the Pentium series, Pentium Pro was introduced by Intel in September 1995, using 0.6 and 0.35 µm technologies containing 5.5 million transistors, socket 8 387 pins, Dual SPGA, support Maximum RAM memory of 4GB. The highlight of the Pentium Pro is a 60 or 66MHz system bus, L2 cache (L2 cache) 256KB or 512KB (in some versions).

Pentium Pro has speeds of 150, 166, 180, 200 MHz.

Pentium II (1997), an improved version from Pentium Pro is used in high-end computers, workstations or servers. Pentium II has L1 32KB, L2 512KB cache, and incorporates an improved MMX technology that makes video, audio and graphics data processing more efficient. The Pentium II has a slot-type socket - Single-Edge contact (SEC) 242 pins, also known as Slot 1.

The first Pentium II processor, codenamed Klamath, was built on 0.35 µm technology, with 7.5 million transistors, 66 MHz system bus, including 233, 266, 300MHz versions.

Pentium II, codenamed Deschutes, uses 0.25 µm technology, 7.5 million transistors, including 333MHz versions (bus system 66MHz), 350, 400, 450MHz (100MHz system bus).

Celeron (1998) is 'shortened' from Pentium II processor architecture, for low-end models. The first version, codenamed Covington, had no L2 cache, so the processing speed was slow, not impressing users. The later version, codenamed Mendocino, fixed this flaw with L2 cache 128KB.

Covington uses 0.25 µm technology, 7.5 million transistors, 32KB L1 cache, 66MHz system bus, 242 pin Slot 1 SEPP (Single Edge Processor Package), 266, 300 MHz speed.

Mendocino also uses 0.25 µm technology with up to 19 million transistors, L1 32KB, L2 128KB cache, 66MHz system bus, Slot 1 SEPP socket or socket 370 PPGA, speeds of 300, 333, 366, 400 , 433, 466, 500, 533 MHz.

Pentium III (1999) added 70 new commands (Streaming SIMD Extensions - SSE) to help increase the performance of the processor in image, audio, video processing and voice recognition. Pentium III includes code names Katmai, Coppermine and Tualatin.

Katmai uses 0.25 µm technology, 9.5 million transistors, L1 32KB cache, 512KB L2, SECC2 Slot 1 socket (Single Edge Contact cartridge 2), speeds of 450, 500, 550, 533 and 600 MHz (100 MHz bus), 533, 600 MHz (133 MHz bus).

Coppermine uses 0.18 µm technology, 28.1 million transistors, built-in 256 KB L2 cache to increase processing speed. Slot slot 1 SECC2 or socket 370 FC-PGA (Flip-pin grid array), with speeds like 500, 550, 600, 650, 700, 750, 800, 850 MHz (100MHz bus), 533, 600, 667, 733, 800, 866, 933, 1000, 1100 and 1133 MHz (bus 133MHz).

Tualatin applied 0.13 µm technology with 28.1 million transistors, L1 32KB, L2 256 KB cache or 512 KB integrated inside the processor, socket 370 FC-PGA (Flip-chip grid array), system bus 133 MHz system. There are speeds like 1133, 1200, 1266, 1333, 1400 MHz.

Celeron Coppermine (2000) was 'shortened' from the Pentium III Coppermine processor architecture, also known as Celeron II, with 70 additional SSE instructions added. Using 0.18 µm technology, there were 28.1 million transistors, 32KB L1 cache, 256 KB L2 integrated inside the processor, socket 370 FC-PGA, with speeds like 533, 566, 600, 633, 667 , 700, 733, 766, 800 MHz (bus 66 MHz), 850, 900, 950, 1000, 1100, 1200, 1300 MHz (100 MHz bus).

Tualatin Celeron (Celeron S) (2000) was 'shortened' from Pentium III Tualatin processor architecture, applied 0.13 µm technology, L1 32KB cache, integrated 256 KB L2, socket 370 FC-PGA, 100 MHz system bus, including speeds of 1.0, 1.1, 1.2, 1.3 and 1.4 GHz.

The Pentium 4 , introduced by Intel in 2000, opened a new era of computer processors and also added "trouble" to users with a number of names, other hard-to-remember plugs. We will continue to 'spot' Pentium 4, Pentium D, and Core 2 Duo processors in the next article and also cannot forget AMD, 'companion' with Intel on the computer 'path'.

Dong Quan