someone (someone.delete@this.somewhere.com) on September 27, 2013 8:25 am wrote:
> David Hess (davidwhess.delete@this.gmail.com) on September 27, 2013 8:06 am wrote:
> > ARM threatens Intel now having followed the same path Intel took through the low end to
> > threaten established microprocessors that held the workstation, minicomputer, and mainframe markets.
> >
>
> The key differences are 1) Intel was a high volume commodity MPU ODM attacking a
> group of incompatible RISC architectures which had far lower volumes and higher
> fixed costs than Intel even if the MORPers were all combined instead of inefficiently
> duplicating each others efforts in chips, system, OS, and programming toolchain
> design; and 2) Intel's designs were comparable and occasionaly higher in CPU
> performance than even the best of the RISCs while being distinctly cheaper.
>
> Although ARM based devices sell in far higher quantities than x86 chips do today
> the utter vast majority of those are extremely cheap devices for embedded control.
> Sales of ARM based application processors/SoCs for phones and tablets are still
> far smaller in revenue than x86 chip sales and a good portion of the profits they
> drive go to OEMs who make the phones and tablets, not ARM or the foundries that
> make the chips themselves. ARM products still can't come anywhere close to main
> stream x86 in performance. Low end x86 have been comparable to ARM in both
> power and performance since CT/CT+. With Bay Trail x86 now beats the vast majority
> of ARM chips in power, performance and Intel is increasingly aggressive on pricing.
>
> The situation today is very much different from that late 80's and '90s'. Simply
> saying ARM will win because it is coming up from below just like x86 did to the
> server RISCs is grossly simplistic as well as almost certainly wrong. The money
> situation is very different and that determines the winner.

While the money involved is relevant, the more relevant issue is the engineering involved.
The flaw with x86 is the insane complexity of the designs. This complexity means
- seven years to design a CPU (and so high latency in responding to changing markets)
- multiple design teams are required (and have to be paid) to create a new design every year
ARM is so much simpler (and will be even simpler when people start making ARMv8 only chips) that one can mount a serious challenge to Intel from below at VASTLY less expense, and on a VASTLY shorter timescale.

The money situation is different, yes, but not all in Intel's favor. To create a competitive chip requires much more expense in labor, and a better quality process. And then to compete requires selling that chip at essentially ARM prices. This is not sustainable. I expect Atom has lost zillions already, and it's going to get worse if Intel have REALLY done what they have claimed (moved Intel to annual tick-tock, with the 5 or 6 or 7 parallel design teams that requires).

Intel have two advantages
- better process
- substantially more experience in turboing a design from cruising speed to bursts of high speed and back.
The second is ephemeral. Apple (and for that matter Qualcomm) have been methodically working their way through the tools of high speed CPUs. They'll get to turboing soon. (Though before then they'll probably do common addressing between CPU and GPU, and HW TM.)
Better process will remain for a while --- but better process costs, and it costs a LOT.

And Intel aren't even just competing against ARM, they are competing against themselves. Every time they make Atom better, they steal market away from their desktop line...
(They've basically learned nothing from the past. What killed Itanium was the deliberate attempt to make it so complex that no-one else could ever copy it --- under the bizarre delusion that that same complexity would not affect Intel's attempts to wrestle with this monster.)