Processors

Intel's 8086 passes the big 3-0

If it wasn't for the 8086, we'd all be still using TRS-80s and Apple IIs. The 8086 CPU turned 30 years old last week. Here's a birthday wish and look back.

The old saying goes: "Never trust anyone over 30." I guess that goes for the computer you're reading this on, because Intel's 8086 CPU had its 30th birthday just last week on June 9. The 8086, of course, is the granddaddy CPU of them all, whether you're running a PC or a Mac. At the heart of all major desktops running today is an Intel or Intel-compatible CPU running 'virtual 8086' mode.

Intel had been producing CPUs for a long time, starting off with the 4004 in 1971. However, within 7 short years, Intel went from producing a 4-bit CPU in the 4004 to creating a full 16-bit CPU in the form of the 8086.

The 8086 was a major advance in CPU technology at the time. When most 8-bit CPUs of the time were limited to accessing 64Kb of RAM, the 8086 could address a near infinite 1Mb of RAM. It also did so very quickly, running at a blazing 4.77 Mhz. In an early version, that translated to about one-third of a MIP. (A MIP is millions of instructions per second -- a key speed rating for minicomputers and mainframes at the time.) Another advantage of the 8086 was that it could use a full 16-bit bus.

If the 8086 had one drawback, it was the price. The first version of the 8086 sold for $360. In 2008 dollars, that translates to over $1,200 -- four times the cost of an Intel Core 2 Quad Q6700. It was so expensive that when IBM went looking for a 16-bit CPU to power its new PC, it went for the lower-cost 8088, which was introduced a year after the 8086 but used a more inexpensive 8-bit data bus.

The 8086 architecture has been declared dead several times. In the mid-80s and early 90s, the argument was between high-performance RISC machines dominating the desktop or CISC processors based on Intel technology. Around the turn of the century, Intel itself tried to drive stake through the heart of the 8086 by trying to force the Itanium line down everyone's throat. Intel rival AMD continued to support and extend the 8086 family line, causing Intel to virtually abandon the Itanium and come back stronger than ever with the Core family.

Dynasties can go on for hundreds of years. In computer years, a 30th birthday is pretty much about as close to a dynasty as one can get.

16 comments
kraterz
kraterz

30 years of being stuck with the short sighted design and market decisions (remember the idiotic 64kb segment / offset memory addressing?) etc. Couple this with microsoft's legacy of lethargy and we've been through the 7 levels of hell. A perfect example of technically superior products losing out to 800lb gorillas with money and marketing muscle. Those of you old enough would remember the Motorola 68000 series was a great leap forward, they totally abandoned backward compatibility and focused on the future, for example, 32-bit registers on a 16-bit processor. It could have made a great alternative to the 80x86 line had they positioned it that way, but we still have the 68000 series around today in microcontrollers and embedded systems.

John Sheesley - TechRepublic Pro
John Sheesley - TechRepublic Pro

A bit of a belated Bday wish for the Intel 8086 in Classics Rock: http://blogs.techrepublic.com.com/classic-tech/?p=142 Intel's 8086 CPU celebrated its 30th birthday last week. Shipping in 1978, it was a big advance in microprocessor technology. The 8086 family has completely dominated the desktop ever since. Even Apple has embraced the technology. Are we doomed to be stuck with ever faster computers based on technology from the Jimmy Carter era?

eM DuBYaH
eM DuBYaH

It was in my Amiga 500!! GAWWD! Tell me that computer didn't totally blow away any Intel/PC of the day!!!

dcolbert
dcolbert

The Motorola 68k was an elegent chip, with great capabilities, but it was slow out of the gate and never caught up. Intel was a *startup* that mainly manufactured memory chips - that was their core Integrated Circuit business, until Japan and other Pacific Rim companies drove them out of the business. In response, Intel began to focus on the CPU, and they were in the right time at the right place to leverage that. IBMs mismanagement of the PC architecture allowed IA86 clones to take off like wildfire. Intel's strong business stewardship, conservative business approach and singular focus over the last 30 years has helped them maintain their dominance in this market segment. All of the (significant) 68k machines were closed architecture. Mac, Atari, Amiga, and their designs and patents were jealously guarded. By contrast, anyone could make a DIY PC and put any IA86 OS on it. This was hugely significant in the success of Intel and Microsoft (although there never was any WinTel duopoly that PC conspiracists like to believe in). Despite Intel's sometimes questionable business practices (and they're only questionable if you're the victim. The share-holders don't seem to mind, as long as they don't get caught. They haven't really been caught often, and the punishment has never offset the rewards, so what incentive would Intel have to behave otherwise?) the IA86 architecture is probably the most significant development for business technology in the last 30+ years. This platform has reshaped society. The way we work, play, communicate, have all been influenced by the emergent dominance of the Intel microprocessor platform. People lose sight that Intel was a ragtag startup that disrupted the business model of business giants at the time. As always, it isn't a matter of technical excellence so much as accessibility. The IA86 processor line is the most accessible CPU ever, and that is the key to success.

Tearat
Tearat

That a 32 bit processor only needed a 32 bit address bus Must have been Intel At least the 8086 a 16 bit processor had a 20 bit address bus Lets see For 32 bit mode 40 bits addressing = 1048576 Megabytes or 1024 gigabytes for 64 bit mode 80 bits addressing = ?? To tired to work it out but it is huge

AressIndia.com
AressIndia.com

That time as the applications were not resource hungry and requirements were limited, 8086 family PC were worked best, even without hard disks installed.

Tearat
Tearat

The business PCs of the time Speed was not the most important thing Compatibility and reliability was But today thanks to MS those are no longer important But speed has now become a business obsession Again thanks to Microsoft Storage what?s that MS and Intel should buy shares in each other I am sure they can achieve a lot more working together In the future we will have massive internet servers that make the calculators of today (computers if you don?t get it) look like the calculators of yesterday (those would be calculators) But they should have just enough power to load what will be the OS of the day in under an hour if you are lucky and treat them well Everyone will know how to backup Insane will be defined by how many new mobile devices you need to look up the bus timetable You will need the bus timetable to plan when to travel to the shop to replace all the obsolete mobile devices that you brought last week because they are the wrong color They are last weeks colour Black black instead of black black black But who knows people may have shot all the salesmen by then Wonder if SteveB will learn how to dance by then The most important question is how will his singing carer be progressing and will he have a job in comedy

dcolbert
dcolbert

It did. They couldn't keep pace, though. And not just a Commodore Amiga issue. It was a fault of the 68k design. It couldn't scale as well. Nothing has. PowerPC, the RISC processors (Sparc, etc). When they talk about Moore's law being broke, as it applies to CPU ICs, it has already broken for several CPU families that couldn't double speed and half price every 18 months. IA86 is the only (consumer grade) processor family that has consistently reinforced Moore's Law (with maybe a minor blip here and there). That is why it dominates.

Tearat
Tearat

That?s not the history I know Read this http://silicongenesis.stanford.edu/transcripts/faggin.htm Intel owes it existence to MS IBM and CBM IBM for their choice of the non standard 8088 MS for the blind obsession toward the elimination of all non-MS standards CBM for the mismanagement of a company that produced the number one selling computer of all time ?IA86 architecture is probably the most significant development for business technology in the last 30+ years.? For business yes if you don?t count the ever increasing costs of IT The price of hardware may have gone down the cost of maintenance in time and labour have gone up For the home there has been much more significant development in entertainment Overall the most significant development is in mobility But the price of oil should kill off most future development in that field The most disruptive force of all time in business will be generation y Just to let you know the Atari ST was based on industry standard chips

Tearat
Tearat

Not just Intels marketing but others Who have a great deal invested in keeping customers locked into the current hardware/software model? Ten points if you can list them all Here is the most well known Intel AMD Microsoft VIA Intel was not interested in dropping its legacy hardware It was only testing the market Well maybe The legacy of the hardware is not the real problem anyway It is the lack of standards in the software You have to remember the PC was such an awesome machine that software emulation of it is totally impossible Just ask Microsoft

dcolbert
dcolbert

I'm not going to undersell the value of good marketing. I still remember ads that showed stacks and stacks of PC apps behind an IBM XT, and a single, small stack behind a competitor's product. But an easily duplicated architecture using off-the-shelf available components didn't hurt things, either. The emergency of the clone/compatible market did far more for the success of the IA86 than any marketing. Ultimately, there were a bunch of factors, and it depends on how you frame the question. The Intel Inside campaign was very expensive and very well engineered. The Bunnyman campaign was also a huge success for Intel. So you can certainly argue that Intel's marketing was a huge part of INTEL's success - but I don't think that necessarily translates into IA86 success. Intel success is about a company, IA86 success is about an architecture that Intel just happened to design. Here is a great example. Although the IA86 architecture could just as easily be associated with IBM as Intel, IBM certainly has had significant struggles since the introduction of the original IBM PC. As far as the 68000, who knows. Fact is, those issues weren't solved. Intel WANTED to dump IA86 for IA64 and their design roadmap called for that. The public wouldn't have it. AMD kept the IA86 architecture viable and alive, and so Intel had no choice but to continue to compete in that arena. I think this is just another example of where if the 68000 had become the dominant architecture, people would be hating it and be lamenting the loss of IA86. The grass is always greener.

Tearat
Tearat

and their decedents dominate because of one thing and only one thing Marketing That?s it and nothing else Changes in the design of other microprocessors would have made zero difference When IBM decided to create the first desktop PC they first looked at buying existing companies 10 points to anyone who can tell me which companies they looked at 10 points if you can tell me what they did not like about those companies Heres a hint The main factors were cost and image The most interesting thing is none of those companies used Intel cpu?s The problems with the 68000 would have been solved The 8086 family still has a long way to go before all the problems are solved The commodore 64 is the highest selling computer of all time Not bad for an 8 bit based 64K system What was its 1 Mhz CPU well it was the jolly old 6502 Not really but the same chip with a small modification It was king in its time The duel core and 64 bit CPU were opportunity?s for Intel and AMD to drop the 8088/6 legacy They decided in their infinite wisdom not to The most significant change in PC design is yet to come Multi CPU data/address bus

Tearat
Tearat

What I find most amusing are the MS fan boys who insist that Microsoft is solely responsible for the modern PC Most of the Intel fan boys have disappeared With XP the choice of office is limited to 2000 till 2003 You will have problems with word 97 and the proofreader on XP But if you use Fat 32 instead of NTFS I think that disappears I have heard that XP and 2007 have problems Then there is ODF Or you can use open office One thing that people forget is you can use OO and MS Office at the same time This is one of the best ways to cut training costs Same hardware, same OS, two choices of office Let them learn at their own pace Don?t forget with MS office the license is between MS and the Owner You are limited on how many PC it can be installed and used on But it is more flexible than the windows license Honestly games are a pain in the ass Most of them are used for a short time then forgotten about But that is entertainment for you Over hyped and lacking in real value If you want to keep the old software running on PC?s Check out DOSBox and any of the other emulators you can find At the moment I am trying out Suns VM VirtualBox That is when I get the time and energy

dcolbert
dcolbert

The thing is, I could still be relatively effective using Office 97 on my XP machine (not so sure about if it will run on Vista). As it is, I *am* running Office 2003, not 2007. For home users without a corporate budget, being able to move their $200 software package with them to newer, fancier hardware is a benefit. Having a piece of software tied to a particular machine and no clear migration path to your new hardware platform is a liability. Now, at some point, you move on, and it all gets thrown out. I get that, and nowhere else is this more clearly seen than in console gaming where emulation or backwards compatibility very rarely have any significant value - more clearly illustrated in the most recent generations of console games. But for business, productivity and entertainment machines in general, software longevity becomes an issue. Wiki was simply the first hit on the google page. There were dozens of other pages. If you want a definitive answer on custom chips in the ST, I'd suggest you check out the forums at atariage.com where you will find some real die-hard Atari Fans from the original pong units all the way through to the Jaguar and Falcon and Flashback 2.

Tearat
Tearat

(ROM) chips were not custom But the contents were The Atari ST was based on industry standard chips There were some custom chips They increased the number of custom chips in later models The main reason was Atari had no money to invest in production The Amiga was a descendent of the Atari 8 bit home computers Which were mostly custom chips With the exception of the 6502C(yes some modifying was done to that) The 6502B was totally standard The other standard chip was the 6520 PIA Atari 8 bit computers were designed for video games The CBM tended to design its 8 bit computers around a single chip Eg the VIC 20 and its video The C64 and the sound device (SID?) That was the golden age of hardware design The skill of the programmers of that day make today?s crop look like amateurs Now its all about the OS and designing around it Be careful with what you read on Wikipedia it is not all true It also leaves out a lot of the more important explanations For the IBM PC The most significant thing apart from open architecture (Which is the sole reason for the existence of today?s PC not software) Was the separate video adaptor Most of the rest of the design leaves a lot to be desired One of the biggest failings of the industry is the development of a standard sound device What is most puzzling is the obsession with legacy (compatibility) and the almost insane need to change everything I can always blame the salesmen anyway It wasn't practical to clone STs because they were the fastest color computer at the time But they were able to emulate the PC and Apples 16 bit computers

dcolbert
dcolbert

The problem is that CBM never had a clear roadmap of a gradual evolution for their line - it was part of the general turmoil of the emerging PC industry of the time. Coupled with mismanagement and bad marketing, it was too much for Commodore to ever overcome. Position was important too. As console gaming grew and waned in popularity over the years, it impacted Commodore significantly. IBM PCs, built on business, had a single front to compete on. Games were secondary and not a driving force for the longest period. A 40 column, 8 bit, serial I/O machine for under $200 was a great boon to introducing people to personal computing, and thus an incredible global sales machine, but all of these things were liabilities as well. The Atari ST, like the Amiga, had custom chips. http://en.wikipedia.org/wiki/Atari_ST "The project, codenamed "RBP" for 'Rock Bottom Price', began to form between April and July 1984 into a design that was almost identical to the ST that eventually shipped. The design was a combination of custom chips and commonly available parts in a highly integrated single-board design, fully equipped with standard and custom ports." TOS on ROM alone basically counts as a custom chip. It wasn't practical to clone STs.

Editor's Picks