cpu guide

In the beginning

Hardly anyone uses these chips anymore, but they were very important at the time — and still are, because all modern CPUs are based in some way or another on these early designs.

This is particularly so for PC users, because the X86 family of chips like the Athlon XP and the Pentium 4 are essentially updated versions of the original 8080 and 8086. Even non-PC designs like the Amiga and the Power Mac are influenced by their history — but PCs are absolutely steeped in it, dominated by it. You must have some understanding of earlier CPU designs like these to understand modern PCs.

Intel 4004

Generally regarded as the very first single-chip microprocessor. It was designed almost by accident: Intel were contracted to create a custom chip for a desktop calculator — they'd done three or four of these already — but decided they could do the job more efficiently with a flexible, multi-purpose chip, which could be re-used again and again in different applications. This was the very modest start of a truly great idea. It didn't power any computers as we understand the term today — the microcomputer was not invented yet — just simpler machines.

The tiny 0.74MHz 4-bit 4004 had a short and unspectacular market life, and very little power even by the standards of the day, but it was the forerunner of the more successful 8-bit 8008 and then the first really successful CPU, the Intel 8080.

Form Design Manufacture Introduction NPU16-pin DIP Intel Intel November 1971 none

Internal clock External clock L1 cache Width Transistor count0.74MHz 0.74MHz none 4-bit 2250

Intel 8080

The classic Intel CPU. This was the heart of most of the world's first microcomputers, including the Altair 8800. (Pictured in the 8085 entry below.) All modern PCs use descendants of the 8080. Indeed, an Athlon or a Pentium 4 can still run something very like 8080 code!

Despite its huge influence and massive sales, the 8080 was soon overshadowed in the marketplace by the more advanced Zilog Z-80, which built on the 8080's success and owed a lot to it. Introduced at 2MHz, as time went by the 8080 was pushed up to just over 3MHz.

(If you want to nit-pick, a dwindling, almost extinct, handful of reasonably modern machines used CPUs that are quite unrelated to the 8080 — mostly the Motorola 68000 family, the DEC Alpha, or the PowerPC. Examples include the Apple Mac, Commodore Amiga and Acorn Archemedes. Strictly speaking, they are indeed microcomputers, but they are not PCs. By definition, a PC uses an X86 CPU.)

Form Design Manufacture Introduction NPU40-pin DIP MOS Technologies Commodore 1975 external

Internal clock External clock L1 cache Width Transistor count2MHz 2MHz none 8-bit 4500

MOS Technologies 6502

A development of the Motorola 6800 by ex-Motorola engineers, the 6502 and its numerous variants powered many of the high-profile early micros, including the Commodore Pet, Apple II, and Commodore 64, not to mention the hordes of marginally compatible Apple clones.

Programmers used to get fanatical about the 6502 because of its simple, elegant architecture. (Old-timers still do.) Even in those days, the computer world was divided into two camps: Z-80 programmers were not good at seeing the subtle power of the 6502 design and felt cramped by it, 6502 fans found the large Z-80 instruction set chaotic and intimidating. Note that the 6502 was very clock-efficient (it was more nearly scalar) and despite clocking only half as fast as a Z-80 or an 8085 it was round about about as powerful. It remained in large-scale production for many years; only the Z-80 and perhaps the 8088 have been more successful.

Form Design Manufacture Introduction NPU40-pin DIP Intel Intel April 1974 none

Internal clock External clock L1 cache Width Transistor count2MHz 2MHz none 8-bit About 5000

Zilog Z-80

The most successful microprocessor of all time.

Zilog was a group of ex-Intel engineers who set out to improve on the 8080 but still maintain compatibility with it. The Z-80 sold in huge quantities, and was at the heart of most of the microcomputers of the CPM era. Apple and Commodore used the 6502, but nearly all the other successful manufacturers used a Z-80. Programmers loved it for its power and simplicity and Zilog pushed it to faster and faster clock speeds over the years. It was more than a match for the clumsy first-generation 16-bit CPUs like the Intel 8086 and can still do useful work. Incredibly, the 25-year-old 8-bit Z-80 is still in production!

Zilog's follow-up designs, however, were over-ambitious, and the 16 and 32-bit Z-800 and Z-8000 were very late to market and suffered from teething troubles. The Z-80, by the way, introduced the idea of SIMD (Single Instruction, Multiple Data) with its block move and copy instructions. These were considered very powerful at the time: real mainframe stuff. Modern 3DNow and SSE instructions work on highly advanced versions of this same basic principle.

Form Design Manufacture Introduction NPU40-pin DIP Zilog Zilog July 1976 external

Internal clock External clock L1 cache Width Transistor count2.5-12MHz 2.5-12MHz none 8-bit with 16-bit elements About 6000

Intel 8085

The illustration on top is an 8080-powered Altair 8800, generally accepted as the very first microcomputer. The switches are the only means of programming it, and different patterns on the LEDs the only output.

The 8085, Intel's replacement for the 8080, had considerable success, but always played second-fiddle to the Z-80. Much improved over the 8080, it powered a fair number of industry-standard CPM machines, especially in the United States.

It was handicapped by an inability to run Z-80 extensions, and by Intel's haphazard programming tools — 8080/8085 assembler mnemonics were very poorly thought out. Most programmers much preferred the dauntingly complex but very logical Z-80. (In those days, all the best programs were hand-coded in assembler for speed and efficiency, and the design of the CPU made a big difference to the way you wrote the program. Today, nearly everything is written in C or C++, which is easier for the programmer but results in vastly bigger and significantly slower software. It is, however, a relatively easy job to port software from one CPU type to another — from X86 to Power PC for example — a task that was near-impossible with assembler code.

In practice, the 8085's main significance was that it kept Intel going until they got their incredibly successful 16-bit 8086 onto the market.

Form Design Manufacture Introduction NPU18-pin DIP Intel Intel 1976 external

Internal clock External clock L1 cache Width Transistor count3-6MHz 3-6MHz none 8-bit with 16-bit extensions 6200

Intel 8086 and 8088

The chip that powered the original IBM PC, and thousands of other models too. All later X86 designs (286, 386, 486, Pentium, and so on) have built on this foundation.

There were four main 16-bit CPUs at this time. The Texas Instruments TI 9900 was an early leader but badly marketed. The Zilog Z-800 was powerful but late and buggy. The expensive Motorola 68000 was generally regarded as the best choice with its combination of simplicity and power. But the 8086 family ended up as easily the most successful of them all.

The main reasons were its compatibility with the 8080/8085 and Z-80 family, its relatively low price, and above all, the 8088 variant. The 8088 was a hybrid 8/16-bit design: 16-bit internally, with 8-bit I/O. This meant that system designers could use cheap and readily available 8-bit support chips, and reduce the cost of the system by hundreds of dollars. Performance wasn't outstanding, and the lean, mean Z-80 and 8085 machines were often superior, but the 8088 sold well.

It got its biggest break when a small sub-division of IBM couldn't afford to pay cash for their first choice CPU, the Motorola 68000, and used the 8088 for the IBM PC. This made the 8088 respectable and the rest, as they say, was history. It sold in countless millions and, as an all time success story, is second only to the Z-80. (Those who claim the reverse order are showing their youth.)

Unfortunately, Intel made some very bad design decisions with this chip. First, it used a segmented architecture (see 286 below). Secondly, for some incomprehensible reason, Intel chose to organise its memory access in such a way that IBM felt constrained to limit the theoretical maximum RAM of the original PC to 640k. Even in those early days, these were short-sighted decisions, and we have all suffered from them ever since. If you've ever had an 'out of memory' error message, then you have met this design fault yourself. DOS, Windows, and even Windows 98 all have base-memory problems which are directly caused by the 640k barrier. Only true 32-bit operating systems like Linux, OS/2 and Windows 2000 escape this problem, and for these we had to wait seven more years for the 386.

Form Design Manufacture L1 cache NPU40-pin DIP Intel Intel, AMD, Harris, Siemens, Hitachi none 8087 or Weitek

Internal clock External clock Width Introduction Transistor count5-12MHz 5-12MHz 16-bit (8086) June 1978 29 thousand4-12MHz 4-12MHz 8/16-bit hybrid (8088) June 1979 29 thousand

Motorola 68000 Family

The best of the 16-bit CPUs, though quite expensive.

The 68000 powered almost everything that didn't have an 8086 or 8088: Apple Lisa, Apple Mac, Commodore Amiga, Atari ST, and many others. The simple, practical design lasted well, and was updated with the 68020, 68030 and 68040. Programmers loved the 68000 (no-one ever loved an 8088) and 68000 code tends to be small, fast and efficient. Notice how 68000 powered Amigas and Macs used to do things in 1MB or 2MB of RAM that you needed 16MB and a 486 to do on a PC.

Like the Z-80 and the 386, the 68000 lives on as an embedded processor for industrial automation work.

It's not meaningful to give transistor count and clock speed figures here, as there have been numerous variations and developments within the 68000 family — much like the X86 family with its 8086, 286, 386 and so on. The last of them were roughly equivalent to the Pentium. By that time, Apple had switched to the Power PC, and most of the other 68000-based manufacturers had disappeared.

Illustration: a 68020 from the mid to late Eighties.

Version Clock Introduction Width Transistor Count68000 September 1979 16-bit 68 thousand68020 June 1984 32-bit .68030 50MHz April 1989 32-bit .68040 25MHz January 1990 32-bit 1.2 million

NEC V20 and V30

NEC were another of the licensed second-source producers of the 8086 and 8088, but they weren't content to simply reproduce: they wanted to improve the breed. (And, of course, improve their margins.) We only ever saw two or three V20s in the flesh, but they seemed to go remarkably well.

The V20 is of interest because it was a faster 8088 (and the V30 the 8086) — the first clone CPU, if you like — and Intel's lawyers were not impressed about it. As they were to do with every X86 chip until just recently, the litigious Intel tried to take it off the market. Eventually, NEC were given the go-ahead by the courts, but by then the 286 was the thing to have and it was too late for either NEC or the computer buying public to benefit much.

There is no need these days to win patent disputes — in fact winning doesn't really matter — it's only necessary to tie your competitor up with delaying actions for a year or two and their market opportunity is gone forever. The pace of change in the computer industry is such that our glacial Nineteenth-Century legal systems are quite unable to arrive at just solutions while it matters, and the bigger, smarter operators are well aware of this. The alert reader will have already thought of several more recent and better publicised examples. These form a recurrent theme which runs right through the development of the modern computer.)

Form Design Manufacture Introduction NPU40-pin DIP NEC NEC, Sony March 1984 8087 or Weitek

Internal clock External clock L1 cache Width Transistor count8, 10, 12 & 16MHz 8, 10, 12 &16 MHz none 16-bit (V20, 8/16-bit hybrid) 63 thousand

Intel 80186

A long-forgotten variant of the 8086 that brought on-board some of the functions of the support chipset. It was not really designed to produce a performance increase, just to reduce cost, size, and power consumption. Not many of these sold as they weren't quite 8086 compatible and the 286 came out soon after. There was an 80188 too. Last time we looked, the licensed AMD version was still made for the embedded industrial processor market — you'll never see one in a PC, but it's nice to think of what's essentially an 8088 at 40MHz!

Form Design Manufacture Introduction NPU40-pin DIP Intel Intel, AMD 1983 80187

Internal clock External clock L1 cache Width Transistor count6-16 MHz 6-16 MHz none 16-bit or 8/16-bit hybrid) ?

286-8

The original 286 (more properly, 80286) was a vast power increase over the 8086/8088, particularly the 8/16-bit hybrid 8088. The 8MHz ones were very rare, as by the time most people could afford to buy 286s they were buying 12 and 16MHz ones.

Like the 8086, the 286 had a segmented architecture — you couldn't just nominate a place in memory you wanted to go to, you had to separately set a segment register and an offset, so there were many ways to get to the same place. This created significant extra work and complication for programmers and had a lot to do with the unpopularity of the world's most popular CPU family. Bill Gates is most famous for saying it, but nearly all programmers called the 286 'brain damaged'.

Gates was actually referring to another less than ideal aspect of the 286 when he said this, but the point remains: we all run variations of the X86 family because they run all our old programs, but even Intel would be hard-pressed to describe the X86 architecture as efficient. The 286 was particularly ugly because although it could access more than 1MB of RAM, to do so it had to switch into 'protected mode', and it can't get back into the normal 'real mode' without a complete reset. In short, it was a terrible kludge.

Oddly enough, the 286 was actually designed some time after the much more powerful 386. While the 386 design was clearly superior, it was just too large and complex and expensive to manufacture at first, and the 286 came into being as a temporary measure — a temporary measure that sold in millions and was in volume production for over ten years!

→ Illustration: a very rare 286 indeed. We can't remember ever seeing a 286-8, though we may well have done at some time, but we are quite certain we have never seen a 6MHz 286 like the one at right, nor indeed heard of one outside a one-word mention in Aad Offerman's Chiplist. In fact, the picture raises more questions than it answers: first, it is the only 286 we can remember seeing in PGA (Pin Grid Array) packaging, though this was to become common with 386 parts and all but universal through the 486 and Pentium eras. Second, if it ever found its way into an AT style system, what speed did the I/O bus run at? (And thanks for sending the picture, Alexei.)

Motherboards for the higher-clocked 286 variants, unlike 8086 and 8088 main boards, had a decoupled expansion bus: the ISA cards, in other words, ran at 8MHz, no matter what speed the CPU itself ran at. This was the first very small step towards the modern multi-clock arrangement, where each of the major sub-systems — mainboard, CPU, RAM, and the various expansion buses — run at their best possible speed and none of them is unnecessarily held back by the larger, slower components they are attached to. It simply wasn't possible to run an ISA bus at the same speeds that the 286 was to become capable of later on: by putting it on a separate clock it became practical to take the CPU and main board to the limit of their potential. Later still, the 486DX/2 would extend this idea to separating the CPU and mainboard clocks, but this was still ten years away.

Form Design Manufacture Introduction NPU68-pin DIP Intel Intel February 1982 80287 or Weitek

Internal clock External clock L1 cache Width Transistor count8 MHz 8 MHz none 16-bit 134 thousand

286-10

Quite a common part, the first of the 286 chips to appear in reasonable volume. Perhaps if we had sampled them while they were still on the leading edge we would have seen things differently, but by comparison to the much smaller and more practical 286-12 mainboards that came along a little later, they were very slow. A good XT was not disgraced by one. (The only one of us who was out of short pants in 1985 was still a happy Z-80 user long after the 286 was old news.)

→ Illustration: as you can see from the unidentified but typical motherboard at left, most main boards that shipped while the 286-10 was selling were complex, cumbersome things with a very high chip count, and few of them survived in working order for very long. At the top left is the RAM: 36 individual 64kb and 256kb chips, making up 640KB in total (counting parity); the two white-labelled chips at centre-left are the Award BIOS; the mid-sized square chip is the CPU itself, and the big square VLSI chips in sockets make up the mainboard chipset — no less than five of them! At far right is the Phoenix keyboard controller.

Form Design Manufacture Introduction NPUUsually 68-pin DIP Intel Intel, various others February 1982 80207 or Weitek

Internal clock External clock L1 cache Width Transistor count10MHz 10MHz none 16-bit 134 thousand

286-12

These were very common, and not bad goers at all. Notice the second-sourcing; Harris, AMD, and Siemens all made them under license. Motherboards had improved a lot by the time these became common: simpler, more highly integrated, and much more reliable.

→ Illustration: an unidentified, fairly late model 286 main board from the days when the 286-12 was the volume seller. There are several things to notice about it: first, the much lower chip count — those three large Headland chips you can see (one almost out of shot) are pretty much all there is to it, aside from RAM and BIOS chips. Next, the CPU itself is AMD manufactured: AMD had yet to venture out into designing their own X86 CPUs but were manufacturing under license from Intel in considerable volume. Third is the empty 80287 socket on the right. Though one was always provided, it was very rare to see it utilised.

Form Design Manufacture Introduction NPU68-pin DIP or PGA Intel Intel, AMD, Harris, Siemens 1983 80287 or Weitek

Internal clock External clock L1 cache Width Transistor count12MHz 12MHz none 16-bit 134 thousand

286-16

The definitive 286, and easily the most common one. We still used to see them from time to time right up to the end of the century, plugging away at DOS wordprocessing, running cash registers, still going strong. The one illustrated was manufactured by Harris Semiconductor, one of the several second-source makers.

This was the 286's high-water mark; they were almost universal: the recipe was 286-16, 1MB RAM in DIPPs or SIPPs or sometimes SIMMs, a 256k VGA card and a 40MB IDE stepper hard drive from Miniscribe, Seagate or Western Digital.

(A short PS: we last re-wrote this entry in the evening of 7/7/98 — and fate being what it is, guess what came into the workshop for an upgrade the very next day: an immaculate AMD 286-16; Headland chipset, 1MB RAM, nice little 40MB Western Digital Caviar hard drive, Paradise 256k VGA card. It seemed a shame to gut it. The system soon became a 200MHz C6 WinChip with 32MB of RAM and a 1.7GB hard drive. We threw the 286 board away, but not without regret.)

Form Design Manufacture Introduction NPU68-pin DIP or PGA Intel Intel, AMD, Harris, Siemens 1983 80287 or Weitek

Internal clock External clock L1 cache Width Transistor count16MHz 16MHz none 16-bit 134 thousand

386SX-16

The 386 was a huge advance but you'd never know it from one of these — they were usually out-performed by the better 286s. Technically, it's possible to run Windows 95 or 98 on an SX-16 — we've seen it done — though the proper term is probably not 'run', it is 'crawl'.

The original 386 (after the debut of the 386SX Intel renamed the original to 386DX) was a full-blown 32-bit monster, and very expensive. This, the 16-bit/32-bit hybrid 386SX, came along three years later. Just like the 8/16-bit hybrid 8088 a few years earlier, the 386SX was designed to provide a way of using existing motherboard components. The SX chip was much cheaper than the DX (though still quite dear) and boards for it, being being only 16-bit and essentially the same as 286 boards, were cheaper too.

Most SX-16s were shipped with 1MB of RAM and were used to run DOS applications like Word Perfect 5.1 or Lotus 123, and they did this quite well. Judging by the retailer's warranty sticker (just out of shot) the one illustrated became some wealthy family's pride and joy for Christmas 1991. This was the time when the Intel stranglehold on CPUs was just starting to break: the AMD 386DX-40 was in production but not yet common and Cyrix were hard at work on their own in-house designs. Before another year was out, the average price of a CPU would have halved and the performance more than doubled.

Even with 4MB or 8MB RAM, you wouldn't want to run Windows 3.1 on an SX-16 though. The SX-33s and DX-40s that followed soon after were vastly faster.

Form Design Manufacture Introduction NPU132-pin PGA Intel Intel June 1988 387SX

Internal clock External clock L1 cache Width Transistor count16MHz 16MHz none 16/32-bit hybrid 275 thousand

386DX-16

Originally just called the 386-16. This was the first 386, and is very rare.

The Intel 386 was easily the most significant X86 CPU of them all. Essentially, modern chips like the 686, the K6 and the Pentium 4 are just very fast 386s. Modern software uses the 386 architecture, the newer chips just do much the same things only faster. The 386 was a huge advance over the 8086 and the '286. In fact, it was designed before the 286, but was too difficult to make and use at first. Intel had to wait and let the technology mature for a few years before the world was ready for the 32-bit 386.

The 386 had vastly better memory management than the 286, and built-in multi-tasking features to allow the development of the powerful modern operating systems that we still use today under various names: OS/2 (which became EcomStation), Windows NT (whch became Windows XP) and Linux. <.p>

Unlike the 286, it could switch back and forwards from Protected Mode to Real Mode easily, and it introduced a wonderful new mode called Virtual 8086 Mode. It is this Virtual 8086 Mode that sets the 386 apart: the chip can run multiple different programs all at the same time, and each of these programs 'thinks' it has a complete 8086 all to itself. This is how OS/2 achieved its incredible crash-protected DOS multi-tasking. Windows NT, and to a lesser extent, Windows 95 followed suit. Virtually all modern software requires a 386, and none of the later X86 CPUs have made the same sort of quantum leap into the future that the 386 did. It was Intel's finest achievement.

Form Design Manufacture Introduction NPU132-pin PGA Intel Intel October 1985 387

Internal clock External clock L1 cache Width Transistor count16MHz 16MHz none 32-bit 275 thousand

286-20

Very rare, very quick. AMD and Harris made these little-known racehorses after Intel decided it could live without its former partners and refused to license them to manufacture 386 parts. Harris went on to do their 286-25 (below), then exited the CPU making business.

AMD took a different view of matters. So far as they were concerned, they had a contract and they were sticking to it. They went on to make their own Intel-designed 386 parts, and eventually were to follow Cyrix's lead and do their own design work too. This was just the very modest start for AMD as a major CPU manufacturer. For the rest of the AMD story, read on.

We didn't do much formal benchmarking in those days — most of the time it just wasn't needed as you could tell two different speed grades apart at a glance — but we always had a strong suspicion that a good 286 was actually faster than a 386SX at the same clock speed, certainly for DOS applications, which was what everybody ran in those days. We should have measured some while they were still available. If we get an old 286-20 come in one quiet day, we'll take a 386SX-25 to compare it to and try it out.

Form Design Manufacture Introduction NPU68-pin DIP Intel AMD, Harris, Siemens About 1990 287

Internal clock External clock L1 cache Width Transistor count20MHz 20MHz none 16-bit 134 thousand

386SX-20

These were a little faster than an SX-16. But then, so was custard.

We vividly remember how they seemed to be slower than a 286-20 too, which they had no right to be: but this was probably more a matter of psychology than of science. After all, you expected a 386 to be fast — it was the very latest technology after all — and you expected a 286 to cruise along at a more modest pace, because it was an old design and slated for the axe before too long. So when you ran them side by side, the 286 perhaps seemed faster than it was. Actually, there was not much difference: both parts could only run the software that was available at the time, and that meant that for all their fancy features they were both operating as mere fast 8086s.

Like all the earlier X86 chips, the 386SX had provision for an external floating-point maths co-processor. For an 8088 you could buy an 8087 co-pro; for a 286, a 287; for a 386, a 387.

The purpose of the co-pro was (and still is) to do complex maths. A standard CPU like a 386 was already very good at doing integer (whole number) maths (at least by the standards of the day), and the vast majority of computer tasks involved simple integers. When some fractional calculations were required, the 386 (or 8086 or 286) was quite capable of doing them too, though more slowly. But some uses require very many complex fractional calculations — examples are scientific and engineering work, CAD, and more recently 3D graphics. The co-pro (or NPU) is specifically designed for floating point maths, and is typically eight or ten times faster at it than the stand-alone CPU. Nearly all 386 boards had an empty socket for an NPU. Current CPUs, of course, all have the NPU integrated into the main chip. (In fact, most computer tasks are integer-based even today — although the role of the NPU has increased over the years, it is still much less important than the integer unit.)

Form Design Manufacture Introduction NPU132-pin PGA Intel Intel January 1989 387SX

Internal clock External clock L1 cache Width Transistor count20MHz 20MHz none 16/32-bit hybrid 275 thousand

386SX-25

We used to think these were a high-performance part once. It seems hard to believe now — it was only by comparison to a 286 or 386-16 that these were anything special.

→ Until the AMD parts brought competition to the market, the 386SX was very hard to justify in price-performance terms. The big price drops that followed made the SX much more attractive for six or twelve short months until the long, long reign of the 386DX-40 began.

In reality, the 386SX did not last well. It began as a cheap way to gain access to the 386 instruction set without having to go to the expense of a 32-bit mainboard, but it soon became an excuse for Intel to shake off the second source manufacturers and charge inflated prices for a rather ordinary CPU.

Although it was substantially dearer than a 286, it was just as crippled by its slow 16-bit memory access. It did offer the benefit of the 386 instruction set — which was vastly better than that of the brain-dead 286 — but it didn't offer any particular performance advantage. By the time software packages arrived that could actually use the 386 instruction set, they demanded performance that was far beyond the ability of even an SX-40, let alone an SX-25.

Form Design Manufacture Introduction NPU132-pin PGA Intel Intel, AMD January 1989 387SX

Internal clock External clock L1 cache Width Transistor count25MHz 25MHz none 16/32-bit hybrid 275 thousand

286-25

We only ever saw two of these. They flew! We had one in the shop for a while, and used to ask friends to guess what was inside it without looking. Most people thought it was a 386DX-40.

We saw an 8088-12 once too. (Actually a V20.) The old XT had a fixed frequency ISA bus, in other words, the video and controller cards had to run at the same speed as the CPU. (In a 286, or a Pentium II, the ISA bus runs at 8MHz no matter what speed the CPU does.) This hot-rod XT used to have to boot at 4.77MHz with the turbo switched off, otherwise the CGA video card would hang! Looking back, we wonder what you were supposed to do if you wanted a maths co-pro chip for a 286-25 — there were several 20MHz 287 chips available but so far as we know nothing at 25MHz. Perhaps you just got a 386 in the first place.

Form Design Manufacture Introduction NPU

68-pin DIP Intel Harris 1989 externalInternal clock External clock L1 cache Width Transistor count

25MHz 25MHz none 16-bit 134 thousand