Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 20 May 1998 16:04:42 +0930
From:      Greg Lehey <grog@lemis.com>
To:        jgrosch@superior.mooseriver.com, Wilko Bulte <wilko@yedi.iaf.nl>, freebsd-hackers@FreeBSD.ORG
Subject:   Re: Original PC (was: talk (fwd))
Message-ID:  <19980520160442.U20476@freebie.lemis.com>
In-Reply-To: <19980519232435.A3703@mooseriver.com>; from Josef Grosch on Tue, May 19, 1998 at 11:24:35PM -0700
References:  <199805191808.UAA17299@dorifer.heim3.tu-clausthal.de> <199805192157.XAA04150@yedi.iaf.nl> <19980520144300.M20476@freebie.lemis.com> <19980519232435.A3703@mooseriver.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 19 May 1998 at 23:24:35 -0700, Josef Grosch wrote:
> On Wed, May 20, 1998 at 02:43:00PM +0930, Greg Lehey wrote:
>> On Tue, 19 May 1998 at 23:57:14 +0200, Wilko Bulte wrote:
>>> As Oliver Fromme wrote...
>>>> In list.freebsd-hackers Mike wrote:
>>>>> [...]
>>>>> I've always heard (I have no motorola experience, yet) that motorola asm
>>>>> blows x86 away when it comes to efficiency.  A friend I have develops for
>>>>> Be and he's always ranting about it. :)
>>>>
>>>> He's right.  The x86 has 4 general-purpose registers, each of
>>>> them 16 bits (they were extended to 32 bits in the 80386) and
>>>> 4 address registers of the same size.  And there are certain
>>>> restrictions on their usage, e.g. you can only use the CX
>>>> register as counter in the "loop" instruction etc.
>>>>
>>>> On the other hand, the Motorola 68k has 8 general-purpose
>>>> registers of 32 bits and 8 address registers (also 32 bits).
>>>> There is no restriction on their use, except that the 8th
>>>> address regsiter is the default stack pointer.
>>>>
>>>> I programmed on both architectures in assembler, and I have to
>>>> say that the 68k is definitely easier to program, and the
>>>> higher number of registers allows for efficient programming.
>>>>
>>>> Maybe it was the biggest mistake ever made in computer history
>>>> when IBM selected the 8088 for their first PC back in 1979.
>>>> (Or was it 1978?  Don't know, I probably couldn't even spell
>>>> the word "Computer" correctly back then.)  If they used the
>>>> 68000 -- which was already available at that time -- we would
>>>> have less problems today, I guess.
>>>
>>> An attractive (to me ;-) explanation is that IBM did not want to use the
>>> 68K because it was a threat (performance wise) to their high profit machine
>>> range.
>>>
>>> Urban legend or not, it sure sounds OK ;-)
>>
>> I just realized what this thread was about after deleting a lot of
>> messages unread, so if I repeat something that has already gone past,
>> please forgive me.
>>
>> Did it really take IBM so long to develop the PC?  My recollection was
>> that it was a sort of half-hearted effort after the devastating
>> success of the 5100.  In any case, the obvious reason for the choice
>> of processor was the software available--CP/M 86 and 86-DOS for the
>> 8088, nothing for the 68K.  On top of that, the 8088 was cheaper
>> because it had 8 bit memory (remember that most chips in those days
>> were single bit).  I don't think they had the slightest concern about
>> attacking their mainframe machines, which were as fast as they needed
>> to be (quite literally).
>>
>
> If memory servers me right, the original IBM-PC was a semi rouge operation
> based in Boca Raton, Fl. I remember reading in Byte at the time, back in
> the days when Byte was worth reading before it became just another media
> toady for Microshit and this is covered in "Revenge of the Nerds" or
> "Accidental Empires", the suits at IBM were very nervous about missing the
> boat with the PC. This is 1979-1980. The Apple II was selling like cold
> beer at a double header in the middle of august all thanks to VisCala and
> Steve Jobs. The comment was "If we do this like a normal (IBM) product
> it'll take 4 years to ship."
>
> I think CP/M 86 and 86-DOS was after the introduction of the
> IBM-PC. 

No, that's not the case.  86-DOS was in fact the name under which
Seattle Computer Products marketed their QDOS (Quick and Dirty
Operating System).  I bought a copy in about November 1980, along with
a couple of S-100 boards.  Does Thunderboard sound right?  IIRC it was
an 8 MHz 8086 and a combined 256 kB memory board [in my case,
populated with only 64 kB] and UART, and it supported *up to 1 MB of
RAM*.  By comparison, the PC looked a little feeble when it was
announced the following year.

> I think the reason they went with Intel instead of Motorola was
> Intel told them the chip was ready, and Motorola told them 6
> months. Intel lied and shipped late. Motorola shipped when they said
> they would but by that time they had missed their window.

No, I don't believe that.  The 8086 had been out for years, and I'm
sure I saw 68Ks in 1980.

Quite honestly, I think that IBM made the correct choice of hardware
for that machine.  Hardly anybody realised what a lasting influence it
would have on the market, and getting the thing out the door was more
important.

Greg
--
See complete headers for address and phone numbers
finger grog@lemis.com for PGP public key

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980520160442.U20476>