[Simh] Why the 4004 ney Intel*64 still lives

[Clem Cole]

On Thu, Jul 9, 2015 at 4:29 AM, Timothe Litt <litt at ieee.org> wrote:

> I agree that IA is the exception - certainly not on its technical
> merits, but rather on its inertia.
>

​Not inertia - economics.   It was a classic HBS (Christensen) style
economic disruption where the lesser (worse) technology (x86 vs Vax/68K
even MIPS) was useful for a different market (PC/DOS later Winders vs
VMS/UNIX etc.) but because that market grew so fast it eclipsed establish
technologies that were based on technology that people described as
"superior."  The the new market does care, the "lessor" technology is good
enough for their use and its the new use that gives them an
economic incentive for that product to take take off.

I remind my peers of this fact all the time.   Don't get cocky about how
"bad" xxx (insert your favorite less technology) is.   IMO: Economics is a
better predictor that technology superiority.   i.e. Alpha would have "won"
if that was true :-)

So it will be interesting to see if Intel*64 can survive the current
"attack" from low-end.  Certainly the wins in the mobile market has allowed
that processor to make huge gains and is clearly make the economics of ARM
interesting.

That said, the lasted European ARM HPC system  [Mt. Blanc  @ BSC]  had a
very interesting factoid that came out in the last few days.  The team at
Rolls-Royce was using Mt. Blanc for their work and discovered:  "Odroid
[ARM] gives ‘cheapest’ run for a single job but Xeon is 27x for 2x the
energy cost."   Which is what I've been saying locally -> as ARM grows up
and becomes more like Intel*64, physics tells us that it pretty hard and
the ARM designer have to start doing the same things.  i.e. if we all
learned anything from PDP-10, PDP-11, Vax, 68K etc.. as architectures move
up from the low-end and start becoming attractive and adding features (like
vectors), they stop looking as much like their low-end fore-fathers and
start becoming more like the systems they replaced.

But if they new lessor technology can establish a strong enough base when
being a lessor technology does matter, you can disrupt the market.   Wintel
did to the Workstation/Unix, which did it Vax/VMS; the Mini did to the
Mainframe etc...

Be interesting to watch and be part of the next shift.


Clem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.trailing-edge.com/pipermail/simh/attachments/20150709/827766c7/attachment-0001.html>