In terms of affordable, general purpose CPUs IBM ARE behind Intel. The Core architecture is fast and cheap, perfect for consumer computers.
Cheap? Then why did the Mac Mini go up in price by hundreds of dollars? Intel is certainly not cheap. Dig up the price lists or just do the math. Intel has the larger chip, and they've got the brand name premium.
The CELL is an excellent CPU for very specific tasks, but very poor for general purpose computing (which is what really matters in a Mac). A CELL equipped Mac wouldn't be nearly as fast as a C2D or even CD one for most tasks people use them for. The CELL is also bigger and hotter than the C2D, I doubt you'd see one in a notebook which is the architecture nearly all Macs are based on.
Sorry, what's special purpose about the Cell other than its first application?
Nothing you're saying here makes much sense. The clock rate race is over because we've hit the wall. Parallel execution is how we'll be getting our speed increases from here on out. Applications will get faster by breaking themselves into lots of little threads that can be parallelized. Why would I want to throw a full fat CPU at each of those little threads? Much better to throw it at a much slimmer processing unit. That will make the device smaller and cooler.
Mark my words, within 5 years and probably sooner, we'll hear AMD moving to heterogeneous cores and within a year or two after that Intel will follow. They'll give it some new acronym to make it sound like a new innovation (VLIW or EPIC?) but it'll be pretty much what Cell is doing now with some mistakes fixed.
If Intel CPU's were that bad, then it shouldn't have been so hard for IBM/Motorola/Freescale to keep up. Not updated as often is different than no mobile G5 CPU after a long, long time. If PowerPC was really lower power, then why did 'G5 PowerBook' become such a long-running joke here?
Most of the heat in the G5 came from the surrounding subsystem. Work the numbers. Driving a few hundred signals off chip at a couple GHz will burn lots, and lots of power. The PowerPC itself was quite efficient. The G4 is phenomenally efficient.
I never understood why people wanted a G5 PowerBook. It never made sense. Do you have 6GB of RAM in your Macbook right now? What we wanted was a multicore G4 style design. Motorola was busy self destructing and when Freescale spun off they decided they weren't going to chase the process curve anymore they were going to stay a couple steps behind and cater to the embedded market.
At the time of the decision, Intel didn't have a competitive laptop processor either-- just a roadmap. IBM was demonstrating lower power G5 variants, but the decision had been made. I'd wager IBM could have met Intel's release schedule on Core and Core2.
No, but my MacBook kicks butt over my old G4 PowerBook.
And in 4 or 5 years your new computer will kick this one's butt. What's your point? I'd argue that the 4 core G4 you could fit in a similar heat profile would do just fine these days...
It's not that I'd argue that Intel is so cutting edge. But they can't be so far behind, if they are able to offer signficantly better CPU's for Apple's family of computers.
Let me put it this way-- take the amount of money Intel burns on their R&D and distribute it among 5 qualified startups. Where would the better CPU come from? I'd bank on one of the startups with 1/5 of the resources. That doesn't mean they'll succeed though because Intel has the market.
And make sure you compare things at the same point in time. Intel CPUs seem better today, but remember Moore expects the number of transistors to double every 18 months and it's been at least that since the last true G5 revision and may be closer to 36 months by now. I think the performance curve for Macs will show little discontinuity at this point if you back up and look at the big picture.