[Ham-Mac] MacOS X x86 Intel Support

[Steve Byan]

On Jun 7, 2005, at 6:12 AM, Richard Rucker wrote:

>
> On Jun 6, 2005, at 8:42 PM, Bill Coleman wrote:
>
>> One curious nit -- no where in any of the PR material does it say 
>> "Intel x86" processors. It just says Intel processors.
>>
>> It could just as well be Intel 64-bit processors, or an Intel 
>> processor of specific design for Apple.
>
> I'm sure you are right.  There is a lot of cruft built into the x86 
> line left over from the early days, or so I have read. The fact that 
> no specifics on the CPU family to be used were discussed I find 
> encouraging.  It means that SJ and Intel have at least one more 
> surprise for us in due time.
>
> I read that some of Motorola's best chip designers on the PowerPC were 
> hired by Intel some years ago.  It seems reasonable that there's a new 
> generation of chips that combine the best of both design worlds; I 
> imagine that would include a high performance-to-power ratio, 64-bit 
> registers, dual-cores, etc.  Perhaps also some features that minimize 
> the pain of low-level code writers where the differences between chip 
> architectures make a difference.
>
> Waiting for the other shoe to drop,

I don't think there's any chance that Intel will produce anything other 
than standard x86 processors for Apple. These days the design costs for 
processors are too high for it to be economical to produce anything 
special for Apple. That's one of the problems IBM and Apple have with 
the G5 - in .13 micron and 90 nanometer technologies, you have to 
choose between fast, high-leakage (and hence high-power) transistors, 
or slower, low-leakage transistors. The extent of the leakage in .13 u 
and 90 nm caught eveyone by surprise - it was much higher than 
anticipated. So now-a-days, you can't just slow the clock frequency 
down and turn a fast desktop CPU into a laptop CPU - you have to do two 
chip designs instead. Apple doesn't do enough volume for IBM to justify 
IBM investing in the development effort for a laptop version of the G5.

Of course the other problem Apple has with IBM is that IBM couldn't 
deliver the target clock frequencies on the G5, even with an extra year 
to do it. IBM screwed up either the chip design and/or the 
semiconductor process.

So the killer micros competition has killed off another evolutionary 
branch; SPARC, MIPS, Alpha, and now PPC are gone and the x86 and x86-64 
architectures are the only surviving lines. I'm sorry to see that the 
technically-inferior design (just like VHS vs Betamax)  won again, but 
that's the way of the network-effect in technology.

Thankful that I've managed to avoid having to deal with x86 assembly 
language,
-Steve
--------
Steve Byan <stevebyan at mac.com>
Littleton, MA