DOCUMENTARY: Why and How IBM ended up creating the PC (and ended up choosing the 8088 CPU)

[Music] [Applause] [Music] hello and welcome along to Al's geek lab in this episode we're going to look into the original PC architecture firstly a we birth the background we're gonna go down memory lane and figure out all the decisions that were made to get to where the IBM PC was launched back in 1981 but we're also going to ask the question why did I VM choose the 808 processor it was controversial certainly it was underpowered indeed it could only address one megabyte of memory which is probably one of its biggest limitations so why did he choose this when other contemporary processors out there who could have been a better choice stay tuned the answers will become clearer the story that led to IBM making the choices in the design of the personal computer are deep rooted in IBM's history the PC marked a huge shift in the way the firm operated so it's worthwhile getting to know a bit of the background on just who IBM were before they made the personal computer incorporated in 1911 IBM pioneered computer technology in the 1960s they brought about the system/360 a revolution in mainframe computing with only a handful of competitors such as Digital Equipment Corporation Dec and HP by the 1960s IBM were almost unstoppable their big blue banner emblazoned across terminals everywhere IBM were a household name especially visible by 1969 when they were publicly seen as a major contributor underpinning the computer technology of their Polo space program with NASA by the 1980s IBM represented over 60% of the global mainframe computer market in addition to their computer equipment they introduced many other firsts including this electric typewriter floppy disk the first practical hard drives and many other firsts the first that they really missed out on though was the personal peuta IBM was making a killing on mainframes and mini-computers that were computers the size of a cabin or two as opposed to a full room they thought that computers at home would remain a hobbyist diamond IBM believed that at work people would continue to use the larger systems as a central computing point using remote terminals on office desks that would communicate with the mainframes rather than standalone computers they were wrong by June of 1977 a fledgling company based in a garage in California called Apple Computer had just come out with the 8 V Apple 2 personal computer it sported color graphics sound heat expansion slots a proper keyboard and was an instant success at an affordable price it revolutionized the market overnight competing in the market very shortly after was the Tandy Radio Shack trs-80 computer the Commodore pet and then later the Texas Instruments 99 Atari 404 and 800 as well as the sinclair zx80 and many other ones by the time the IBM got round to thinking about a personal computer the market was already thriving worth at least a hundred and fifty million dollars in 1979 and expected to grow by 40% in the early 1980s IBM were clearly on the backfoot they needed to play catch-up fast before the rise of the microcomputer IBM was known as the safe bet in computing a well-known phrase nobody ever got fired for buying IBM is the thing of legend because there was a growing market in home and office microcomputers an IBM weren't present in this sector businesses were left scratching their heads they were buying from companies such as Commodore and Apple instead of the safe bed IBM were not happy not only did I BM not see the microcomputer revolution happening when they finally caught and on some key staff for IBM knew that the wheels of inertia was so slow at developing a system that could compete in the microcomputer market would mean that the market could pass them by altogether the truth was IBM were embroiled in red tape and management hierarchy as well as a solid research and development process IBM were the epitome of proprietary they would develop everything themselves in their own labs processors memory motherboard design none of it was off-the-shelf components all of this was great for quality control and solid products but really bad when it came to developing a new product quickly IBM didn't have the agility of companies like Apple that could create a computer with a couple of guys in a garage in a matter of months in fact when asked for an estimate on how long it would take IBM's engineers to develop a microcomputer using IBM standard processes their engineers said it would take them four years to produce a product the year was 1980 there had been at least three to four years steady growth in the microcomputer market which IBM wasn't in at all [Music] an IBM executive by the name of bill law was one of the staff at IBM who foresaw the rise of the microcomputer in fact he had started working on trying to sell personal computers to IBM as far back as 1973 by the time 1980 rolled along Bell tirelessly tried to convince senior staff that in order to enter the microcomputer market IBM would have to write at leat Inge the way they work if they were going to get a product out in time finally in September of 1980 the chairman Frank Carey and then CEO John Opel agreed to make a skunkworks division of sorts to rapidly develop a personal microcomputer despite IBM employing over 340,000 staff in 1980 Bell was given a team of just 12 staff to develop the personal computer or simply PC as it became known in time so the skunkworks began in earnest one of the early options was to acquire Atari and base their new machine on the Atari 800 computer 800 shipped in late 1979 and used the maus technology 6502 processor an 8-bit CPU acquiring an outside company such as Atari was a move that was practically unheard of in IBM history it was hardly a surprise then that IBM executives rejected the idea given that an acquisition was not an option Bell took his team to task they saw that the best way to bring a decent microcomputer to the market was to use mainly off-the-shelf components and design the machine with an open architecture this was very much against the design philosophy of IBM but it was clear that if they were going to get muck to market in time research design and creation of their own system using proprietary IBM hardware was simply not going to work when IBM's executives agreed that Bill's plan was solid in time they allowed up to a hundred and fifty engineers to join a team of twelve they had just 12 months to get it shipped the open architecture allowed the PC to connect to peripherals and expansion cards made by third parties to the machine IBM knew that they wouldn't have enough IBM branded peripherals ready on day one they were also following in the footsteps of Apple the Apple 2 is eight expansion slots were a big part of the reason that the Apple 2 was so successful many of bells team were also Apple 2 owners at the time and were quick to raise this advantage with the team bells team figured correctly that the industry would create many more expansion offerings that greatly increased the usefulness of the PC before 1981 was out a booming Cottard industry had already started making all sorts of add-on cards such as speech synthesis cards Mouse interfaces z80 daughter boards RAM cards disk controllers graphics cards you name it these add-ons were coming out thick and fast when the PC shipped IBM received no patents on the PC this is important this was due to the fact that 90% of the PC had been manufactured by third parties indeed the only major proprietary part to the PC was it's warm BIOS without this early decision to make the architecture open who knows if the PC would have had the success at dead as a sidenote years later with the release of the ps2 systems IBM tried to undo the open architecture of the PC and make a new proprietary expansion bus called MCA or micro channel architecture forcing any third parties to pay a license fee to IBM if they wanted to make expansion cards in the end MCA filled opening the way for companies such as Compaq and Dell to make clones of the IBM PC using the extended 16-bit version of the PCs original open expansion architecture amongst the many decisions on what to choose for the new computer architecture the most major would define how the rest of the PC would work this of course is the CPU the brains of the machine the explosion of the integrated circuit and computer on a chip solutions had been threatening the many computer and mainframe markets long before IBM got into the market so by the time the IBM came to choose a processor for the PC there are quite a few options to choose from indeed there were a proliferation of 8-bit processors as well as even 16-bit and some 32-bit processors available the first 32-bit processors came out in 1980 and 1981 one of which was a 10 mega her IDM process are called rompe a sophisticated reduced instruction set or risk cpu running at 10 megahertz and true IBM style despite the CPU being readied in 1981 the first computer that actually shipped with the romp was in 1986 the RT PC had the original 1981 PC shipped with this processor the PC would have been massively ahead of its time opening up the possibility for multitasking and real time operations that were only seen in the lakes of higher-end many computers of the time alas this wasn't to be Fellowes team had a few credible options for processors at the time with 24-bit and 32-bit processors being a bit far fledged it came down to good old 8-bit processors or the more powerful 16-bit processors processors either came with the Cesc complex instruction set or risk reduced instruction set there are pros and cons of each design that waged Wars for many years with Apple and others including IBM at times setting in the risk side of the argument the credible 8-bit processors off the time where most technology 6502 and it's follow-up the 6 5 or 2 C it was first introduced in 1975 it ran at 1 mega ha and was featured in many original microcomputers such as the Apple 2 Commodore pet BBC micro Atari 800 commodore vic-20 and even the atari 2600 console well a variant thereof effectively it was a risk-based processor although it ran at 1 mega her it would often appear competitive in speed with other contemporary CPUs thus was enormous success for many years into the 90 ATS and although simple and design had many fans and later followed up with a 16-bit descendant called the 6 5 C 8 1 6 in 1983 it was most notably used for the Apple 2 GS range of computers at that point in time as with most processors the 6502 could address up to 64 K of RAM there was also the z80 CPU it was a product of another 12 employee sized company made up of people who left until z80 was dialogues first product it launched in 1976 and was also a massive success at me to enter machines such as the Sinclairs 880 and Z 81 as well as the spectrum later in 1982 in the office it was used in more serious machines such as the K Pro Osborn 1 and the IPPS and qx10 computers its low heat output made a successful in the portable embedded and handheld computers industry and is still being used today in many consumer electronics systems interestingly it is a software compatible extension of the Intel 8080 CPU the zet originally came in at 2.5 megahertz and later a revision ran at 4 megahertz with later generations even running at 6 8 and 10 megahertz the CPU could support up to 64 K of RAM other possible contenders could have included the quasi 16-bit motorola 6809 which was used on the trs-80 color computer and the vetrix as well as the 8-bit 808 5 from intel more on that one in a minute though however fairly early on in the development of the PC the Dirty Dozen has they've become known decided that the fully 8-bit processors wouldn't cut it the team had the foresight to say that having a maximum of 64 K RAM was a limitation that wouldn't be best suited for business applications of the future the Commodore 64 came out a year after the IBM PC did and it had a 64 K ram limit so this was not a limitation that other manufacturers felt was so important and many other machines came after that that also sported 8-bit processors with 128 year however they had to make use of the difficult Bank switching memory operations techniques the IBM team decided that the PC should support even higher memory capacities than the 8-bit processors could so from day one the PC shipped with options to house 256 key RAM and entire expansion boards would take it to 640 K without any other system modifications so after the 8-bit processors had been ruled out that just left 16-bit processors firstly the Texas Instruments TMS 9900 a 16-bit single chip version of the TI 990 mini computer or the deck PDP being 16-bit it could only address up to 64 K of RAM it was used in the TI 99 - for home computers Texas Instruments Walden C rhymes gave a presentation to IBM's team that were developing the PC at the time according to reins TI lost i/o and the opportunity because IBM had already received production samples of the Intel 8086 which were going through quality assurance at the time in the mid-1970s Texas Instruments were the largest semiconductor company despite this after a few more attempts at making computers Texas Instruments stopped work on general-purpose micro processors shortly after next up was the Motorola MC 68,000 released in 1979 known as the 68k it was designed from the ground up with assist instruction set without backward compatibility for the previous 8-bit 6800 range computers initial speeds of the processor was epi available at four six and eight megahertz with 10 megahertz 12.5 megahertz and even a 16 point 7 mega Harrison being produced through to the early onlet late 1980s there were instant favorites for UNIX workstations and favorited by companies such as Sun Microsystems despite it being a 16-bit processor I had a 32-bit instruction set and a data bus and a 24-bit non-segmented memory at boss making it ideal for programming it supported up to 16 megabytes of physical memory the memory capacity eclipsed that of contemporary CPUs by quite some amount and just to be clear here the Intel processors that were available at the time supported a maximum of one megabyte ram the 68k was hugely popular and computers that arrived in the mid to late 80s including the Apple Macintosh the Commodore Amiga the Atari ST and many others outperformed the Intel equivalents at the time easily it's still in use in many systems today forty-one years after its first production run that leaves the Intel 8086 I've already touched on the fact that Texas Instruments were rebuked by IBM because they were already trialing the 8086 however in the end IBM chose not to use the 8086 either released in 1978 the 8086 was a true 16-bit processor initially running at five megahertz the 8086 was the first in the line of the long-standing x86 processors with the 802 86 386 and 486 line continuing the trend the CPU had a 20-bit address bus which allowed support for up to 1 mega byte addressable Ram unlike the non-segmented memory bus of Motorola 68k the 8086 allowed a maximum linear address space of 64 K meaning the despite supporting 1 megabyte of RAM you had to write code to using segment registers making programming the CPU a little bit more cumbersome it was not until the advent of 1985 8 or 386 CPU a true 32-bit microprocessor was this design constrained result despite at running at 5 megahertz the CPU had to wait for quark memory access cycles making it slower than contemporary CPUs for 8-bit data transfers but slightly faster for 16-bit transfers there are also some other technical limitations such as a slow instruction fetch in execution units which hindered performance a little as well this is true even with today's Intel and AMD x86 processors later variants of the x86 introduced in 1980 supported speeds of up to 10 megahertz finally we get to the CPU that Kim in the IBM PC around the year after Intel released the 808 6 they came out with the a 2-way eight or eighty eighty a to some knowit it was a variant of the 808 sex the 88 was architectural II very similar to the 808 six the main difference is that the 808 sex is a true 16-bit processor the 808 eight is not internally the 808 eight is a 16-bit processor but the 808 eight has an 8-bit data loss whereas the 808 six can read and write eight or sixteen but later at time the 808 eight can only do 8-bit data transactions at a time the clock speeds of the 808 six were five eight and ten megahertz whereas the 808 kmn are only four point seven seven and eight megahertz although later clones like the NEC v20 CPU supported up to ten megahertz performance wise the 808 eight could run about half of the speed of the 808 six and the only benefit of the 808 eight over the 80286 is that I was fully backwards compatible with the older 808 5 processor whether this was off any real benefit to most everyday PC users is arguable but it might have made more sense at the time due to the prevalence of the cp/m operating system which was intended to ship with the IBM PC rather than ms-dos CPM was typically seen on z80 processor base machines at the tank and the 808 5 was a better performance version of the 8080 processor from 1974 the 8080 was the genesis of the z80 processor and was indeed mostly compatible with the z80 which allowed from much of the z80 software to run on the a to 8/8 without significant changes so if they intel 80286 and the mortal role of 68k were clearly better chips than 808 why in the heck did IBM choose the 808 8 some may argue that it's all down to one or two reasons but I've checked and I can find up to five different reasons why they chose the 808 eight the first one is simple cost the list prices of the initial 5 megahertz of the 80286 on release was slightly cheaper than the 8 or 88 87 dollars versus 125 this was strange because both chips featured exactly the same amount of transistors however when IBM went to buy a CPU from Intel they were offered the 808 eight at a cheaper unit price than the 808 six instead of going down in price the 808 six is newer and faster variants when up in price this is unusual with regards to CPU pricing following Moore's law don't sue me please but something here sounded a bit fishy so yes at the time that IBM got the 808 eight the processor was cheaper than the 808 sex but looking up a bit farther afield choosing the Motorola MC 68,000 on a purely cost basis it was clear that Intel won out in that area the 68k was priced at approximately 350 dollars more than the ADA wait 8 off-the-shelf memory chips where 8 bits at the time so to use a 16-bit CPU and access memory efficiently ie 16 bits at a time you'd have to use twice as many memory chips and memory chips were expensive so that would add up to $400 more to the price so you can see that both this cost of the CPU and the memory would have made the PC untenable compared to other contemporaries the second reason why the eight weighed eight was chosen is that Intel claimed that availability of the eight aware aid was better than the 808 sex at the time of production in 1980 Intel were able to supply more 808 eight chips it is quite probable given that both the 808 eight and the 808 six had 20 nanometre fabrications they were made at the same factory who knows also who knows of Intel had in fact made too many 808 eighths and we weren't selling so they decided to make a BM and offer that was too good to refuse again pure conjecture the third reason is go with what you know or research and development team the personal computers of 1980 were almost all still 8-bit machines if IBM were to make a 16-bit bus it would require researching a whole motherboard design rather than something which was familiar at the time not only would the motherboard have to be quite different from the 8-bit machines they would have to design new expansion slot types that wouldn't necessarily use off-the-shelf components in the end the 8-bit expansion ports that were used on the IBM PC were incredibly similar to those on the Apple to the Motorola 68 km was also not completely proven it had only just come to market the chip wasn't fully debugged or proven as a safe bet by appearing and other personal computers IBM would be pioneering by using this chip and as we know by now beg blue were known for being conservative when it came to risk it was clear that if IBM wanted to use the 808 sex or the 68k they would have to spend some time in research and development compared to having an 8-bit bus schematics for 8-bit machines were widely available with designs by third parties in the public domain this meant the IBM didn't have to do very much to make an 8-bit bus motherboard so looking at it with a pure project management hat on time to research 16-bit architecture costs money and on top of that every moment that IBM didn't release the personal computer was time that the rest of the market could take market share away from IBM the cost of delay was doubly high one of the engineers who joined the PC team David Bradley had just finished work on a machine that was finished around August 1980 known as the data master or IBM system 23 there was a small all-in-one computer it ran IBM's own basic language the idea was it was easy enough for anyone to install in their own office although this was a totally different system and it was part of IBM's long time to release legacy work started on it four years before it was released it wasn't fat a desktop sized personal computer in fact this is one of two machines released by IBM that were effectively personal computers both of which came out before the PC both of them flopped mainly due to price interestingly though in a case of the data master IBM had chosen to use entails a to a5 processor as mentioned earlier the 808 processor had backward compatibility with the 808 5 or z80 CPUs this meant that design principles used on the data master could be reused in some cases on the PC when David Bradley finished work on the data master team he said that him and his team had now become intimately familiar with the intel architecture and support chips lessons learned from the data master also included the fact that designing their own basic language was a very slow process David knew that using an existing basic would speed up this process this is the main reason that Microsoft's basic was embedded into the rom of every PC and XT the data master included the keyboard monitor and system unit in one box the team learned that having the keyboard attached to the chassis of the PC wasn't going to be a good move however they dead reuse the exact same keyboard which became known as the model F the fourth reason that IBM chose the 808 eight was perhaps because of outside influence in a copy of info world magazine from August 1982 Bill Gates was quoted that there was a lot of input into the design of the IBM PC from people at Microsoft this input was made even before IBM green-lighted the development of the PC gates went on to say that Microsoft made influential statements about not using the z80 CPU there was some consideration of having it be a z80 processor one of the key engineers and I held out breath and said look it's gotta be a 16-bit processor the article then states many people wondered why the IBM selected the 8080 processor rather than the 8086 which would have been a more powerful selection to which Bill Gates is quoted as chuckling that was complicated looking back I have to say that in retrospect it was a close call I sure know a lot of cases where it would be nice to have additional machine horsepower the article goes on to say that had IBM chosen an 8-bit processor scheduling of the machine would have been easier because a vast amount of zet software was already available at the time now I'm not suggesting here that Bill Gates deliberately sabotaged the use of the z80 processor in the PC so that he could use an Intel compatible operating system which would become ms-dos but the obvious choice for an operating system at the time was digital research's cp/m operating system indeed this was actually IBM's first choice rather than Microsoft's dos asked later IBM were tight-lipped about Microsoft's involvement simply saying that they went to Microsoft because of their basic interpreter it was very good and widely used IBM spokesman John to Adele said IBM designed their PC Microsoft looked at the design and said that IBM had a quote nice machine the fifth and final item on the list is about software apparently there had to be software applications and more importantly an operating system that was available for the processor Seattle computer products had an operating system known as 86-dos or kudos as it was first known kudos interestingly stood for quick and dirty operating system Microsoft bought kudos from SCP for $75,000 and in turn they licensed it to IBM who rapidly needed an operating system that would work with an 808 6 or an 808 8 digital research's CPM 86 operating system just wasn't released in time whether it was cost to produce availability research and development time outside influences software or a combination of all of the five the choice to put that fear engine in a potential Ferrari was cemented sometime in 1980 whether it was the right choice or whether it could have been pushed a little more is something that we'll never know but what we do know is that later PC clones such as the Amstrad PC one five one two shipped with the 808 six instead of the 88 and performed favorably over the other IBM equivalents all the choices had been made and by the autumn of 1981 the first PC arrived the 5150 the very first computers seemed as babies houses and so mysterious that for most of us the computer was behind a closed door but IBM was thinking how to make the computer more useful and is one good idea led to another it began getting smaller faster less expensive and easier to use today a new IBM computer has reached a personal scale a person can afford it a person can put it anywhere office or school and a person can learn to use it with ease IBM made its personal computer to help a person be more productive to help the person be more creative [Music] and those are good reasons proportion to feel good the IBM personal computer now it selected stores across the country the base model was equipped with 16 K of RAM with that 4.77 mega her 8 to 88 processor a cassette interface nor monitor having a TV interface instead the base model without any extras would cost you one thousand five hundred and sixty-five US dollars in 2020s money that equates to around $5,100 considering that other systems such as the Apple 2 were selling for less than $1300 and a trs-80 cost just $5.99 including a monitor if you wanted a half-decent IBM PC you could upgrade to 64 K RAM a floppy disk drive and a monitor which would set you back $3,000 or approximately 9,800 in today's money the PC wasn't the cheapest system out there by any means but it did have one big advantage thought the sticker on the front seen I beat em by the end of 1982 an IBM PC was sold every minute of every business day so to conclude today's PCs are still effectively the same as that 1981 system they still use the same Intel type architecture and can still even run the same software if pushed to do so despite Apple's resurgence around the mid-2000s and a conversion to using Intel based processors they still only hold a market share of 6.3% IBM sold the personal computer division to Lenovo on the 1st of May 2005 IBM CEO at the time samuel Palmisano stated that the company wasn't the standard setting organization that it had been whilst the personal computer division was still profitable at the time Palmisano didn't see the PC as innovative or a long-term growth market despite many people agreeing that the PC market would slowly declaimed to nothing in the face of new technology including tablets mobile phones and other on-demand consumption methods this hasn't quite happened as people expected indeed in 2019 IDC and Gartner concluded that the PC market actually grew four point seven percent and has been growing steadily since its low point in the last decade and with over quarter of the market sales guess which company is top of the roost in the PC market yep you guessed it it's little despite is relatively high cost it's questionable design choices and it's no-frills style the IBM PC created an industry standard that long outgrew IBM in my opinion the law powered a 288 CPU was a poor choice that defined a legacy of setbacks and compromises that lasted into the 90s fixed by the eight or 3/8 X however unless you are watching this video on a phone or a tablet you are watching on a descendent of the 808 8 despite the poor choice of CPU the PC endured because of many other aspects the IBM badge the open architecture and documentation its expandability the killer applications like motifs one two three one can only wonder how the world will look today if the PC wasn't the platform that OneNote we'd almost definitely not be using Microsoft Windows as it is today maybe not even using Microsoft products at all Apple solid or Britannia logos by even adorned every day store and perhaps a more proprietary existence might have taken shape for me though I'm glad that the IBM compatible PC one the computer Wars of the 80s it unified the computer landscape of the 80s and early 90s and created an open industry standard that we have come to take for granted to this very day thanks very much for watching if you liked this episode and want to see more please subscribe below if you've got any comments I'd love to hear from you until next time take care and have a great day you

Channel: Al's Geek Lab

Views: 133,540

Keywords: IBM, 5150, ibm, Atari, Apple II, TRS-80, 8088, 8086, CPU, vintage, retro, Intel, Digital, mainframe, Bill Gates, Microsoft, ibm pc 5150, the original ibm pc 5150, ibm 5150, ibm pc, computer documentary, retro computers, vintage computers, Classic computers, ibm documentary, ibm history, bill gates, apple 2, vintage computer, vintage computer commercials, ibm keyboard, ibm mainframe, 8088 computer, ibm selectric, retro tech, 8088 microprocessor architecture, ibm pc documentary

Id: -WwWfAKsS4w

Length: 36min 58sec (2218 seconds)

Published: Sun Feb 02 2020