Re: Wozniak vs Jobs

[Jeff Lemke <lemke@terra.cira.colostate.edu>]

Supertimer wrote:

> I have just one comment regarding the above -- you can't compare
> Intel Mhz to Motorolla Mhz to 65C Mhz.  That 1Mhz Apple II could
> often match (or even beat in some cases) the speed of the original
> 4.77Mhz 8088 IBM PC.  It has to do with how many cycles the CPU
> takes to execute instructions.
> 
> In addition to an accelerated Apple IIGS, I own an IBM AT 286 at
> 8Mhz and a Turbo XT.  I timed the Apple IIGS, setting the on board
> accelerated CPU to match the Mhz of the Turbo XT's 8086 and
> found that (1) unzipping files and (2) decoding a JPEG file both
> took much longer on the Turbo XT than on the accelerated GS set
> to the same Mhz.  In fact, an 8Mhz 65C816 is most comparable
> to a 25Mhz 286 in terms of performance (a fact that I can appreciate
> because my IBM AT runs at 8Mhz -- it is the revision 2 model).
> 
> When you say "advocate of Apple II that ran at 2.8Mhz," know
> that when that Apple II's 65C816 ran at 2.8Mhz, it often kept up
> quite well with that IBM AT 286 running at 8Mhz (for example,
> when running the GS/OS Finder compared to an IBM AT running
> the GEM environment...with GEM in CGA or Hercules...moving
> LESS screen data than the GS)...
> 
> (Although MIPS are often as inaccurate a comparison as Mhz
> between different CPU architectures, it is useful to know that
> when the 65C816 is run at 16Mhz, it gets 8MIPS....which is
> better than a 286 at the equivalent Mhz...)


Yeah, I know that MHz isn't a very good yardstick by itself - it is sort
of like trying to determine your car's speed by the RPM your car engine
is running at.  You also have to know what gear you're in.

This point was most vividly driven home to me just yesterday.  I was in
a special travelling exhibit truck of Silicon Graphics computers on
Colorado State University's main campus.  As a graphics artist, I felt
like I was in Tomorrowland.  Silicon Graphics machines run IRIX 5 (a
Unix variant) and use 64-bit R10000 (R-ten thousand) RISC processors
that crank at 250 MHZ, and while at first glance the MHz rating is less
than recent Pentium II's, they simply blaze like nothing I've ever seen
before.  They blow the doors off any 32-bit 300 MHZ Pentium II, or even
any Power PC chip.  It stunned me how they could do real time
interactive 3-D animation on a super wide screen (9:16 HDTV aspect
ratio), very hi-res display.  Not a hint of flicker or dropped frames -
and this is for a resolution of something like 1820 x 1024.  Incredible
for real time.  Now granted, the interactive display was not a super
complex scientific modelling visualization, but it was highly detailed
and something that was in no way possible in real time with any 32-bit
chip, CISC or RISC.  In another end of the truck was a flight simulator
fitted with 3 wide screen hi-res displays put together, all being run as
a single, very detailed terrain fly-by on a 4-processor Octane (one of
SGI's models).  Two years ago they bought out Cray Research, partially
to obtain the technology for ultra-high-speed data bussing for massively
parallel computing, which can allow 128, 256, or even all the way up to
4096 processors to be simulataneously linked into essentially one
computing entity.  (Pardon me if my terminology is less than accurate -
I'm doing the best I can from memory.)  Such computing arrays are used
in governmental and research institutions, aerospace, and in movie
studios in computer special effects and animation.

I got off the track here, but it's amazing what doubling the bandwidth
can do.  It does apply to what was said about the 8088 not being a true
16-bit processor (9-bit buss), and that the 65C816 chip was able to
perform functions with fewer clock cycles.

 

> >Business people were reluctant to begin
> >with to buy any computer systems that didn't have Big Blue's logo on
> >them, but a few businesses did try the Apple III's, but Apple stumbled at a crucial time.
> 
> Apple stumbled in business because the A3 was used to try
> to break into business instead of a new A2.  You have to realize
> that the A3 had A2 emulation, but a deliberately crippled one.
> You could not access the A3's features from A2 emulation.
> That's equivalent to building a computer that prevented DOS
> programmers from accessing the VGA graphics screen because
> you thought they would not use Windows -- while your old
> DOS computer COULD do graphics and you had software that
> used graphics...!


The stumble I was referring to was putting out Apple III's that didn't
work to an already reluctant business community.  Word gets around
quick.  Anything Apple tried to do after that for damage control was too
late.  


Jeff Lemke