[rescue] Apple to ditch IBM, switch to Intel chips

[Patrick Finnegan]

On Sat, Jun 04, 2005 at 03:15:10PM -0700, Devin L. Ganger wrote:
> on 6/4/2005 1:21 PM pat at computer-refuge.org wrote:
> > The fact that you've even mentioned it means that you must think it's
> > still a relevant measure between different processor types...
> 
> It *is* a relevant measure when you are talking about market share, if 
> only because so many people out there *think* it is a useful way to 
> compare processor architectures despite the actual technical truths. 
> People want a single easy comparison that they can tell themselves they 
> understand, and "higher number = faster = better" is about as easy as 
> they come.

Do they?  Maybe I've just isolated myself from such idio^Wpeople, but I
don't hear anyone really talking about clock speeds anymore.  Even Intel
has stopped emphasizing it so much since they released their Pentium M.
AMD makes it hard to determine the actual clock rate as well.  You still
see it on basically all non x86/non AMD processors to demark speed
though.

> Nothing I saw in Chris's response indicated that he thought it was a
> valid *technical* measure, and his point is valid -- Apple has done a 
> very good job of selling systems despite the fact that their systems are 
> based on a chip whose architecture is commonly perceived as vastly 
> slower *by the mass market*.

I hope Chris didn't take my comment personally. :)

Anyways, most people I know don't care about clock speed at all anymore.
Maybe it's because I've just about completely isolated myself from the
WinTel world, or maybe most people just don't care.  I think it depends
a lot on the exact market you're looking at, but most people seem to
want the cheapest machine with the biggest harddrive that you can get
from Dell or Best Buy.  Neither of those sell Mac's, which my be another
reason limiting their popularity.

Anyways, the few people I know who care about speed of machines are
"customers" of where I work, and they are mostly interested in spending
the least amount of money to get their one piece of research code to run
as quickly as possible.  They're generally smart enough to realise that
GHz isn't the only measure of a computer's speed.

Pat