Re: Advice on system purchase
On 10/30/2012 4:16 AM, Joe wrote:
> On Mon, 29 Oct 2012 23:02:58 -0500
> Stan Hoeppner <email@example.com> wrote:
>> No, I mean millions. One billion chips per year would equal 1 for
>> every 7 humans on the planet, and that's simply impossible. Over 3
>> billion people have never used an electronic device. That's almost
>> half the Earth's population. Do the math.
> According to ARM:
> "2.2 billion chips shipped, split equally between mobile and non-mobile
The context of this sub-discussion in this thread is ARM chips that
might be suitable for a desktop PC. That eliminates all the embedded
chips in printers, DSL modems, consumer wireless routers, set top boxes,
etc. So the numbers we're talking about are in the hundreds of millions
tops, as only the chips in high end smart phones and maybe some
enterprise networking gear are possibly suitable.
>> It's simple economics: If one could make a decent amount of profit
>> pushing an ARM based desktop CPU into the market, they'd do it. They
>> haven't done it, nor will do it, because there's no money to be made,
>> only losses, as history has shown us. Both IBM/Motorola and DEC lost
>> money and failed to drive adoption of their RISC chips in desktops.
>> Apple dropped PPC for Intel, eliminating the last RISC CPU in desktop
>> machines. Given this history, if you're an exec at ARM, would you
>> consider such a push viable? Let alone profitable? No, you wouldn't.
> There once was money in it, and they did. The first standalone ARM PC
> saw the light of day at the end of the 1980s, when it was faster than
> the 386s of the day and competitive in price. It also ran a useful
> desktop OS when MS was still fumbling with Windows 2.0.
The ARM based Acorn computers were not a commercial success and did not
produce meaningful cash flow or profit.
> They just couldn't keep it that way, they had the Apple trouble of
> being a single manufacturer of computer, OS, and in this case, CPU.
That wasn't Acorn's problem at all. Their problem was the same of their
contemporary's. Back then, unlike today, there were dozens of companies
with proprietary processors and software competing in a space with no
software standards. The IBM PC sucked compared to many of these
machines that offered real innovation, but if offered at least some
semblance of a standard, and had far more ISV support than any other
platform. When MS Windows matured and ISVs coalesced on the platform,
it sealed the fate of the non x86 machines, and even some x86
(8086/8088/80186) machines that weren't IBM PC compatible.
> This kept the prices very high, but not for long. Development was
> far too slow to keep up with the rest of the world. Even Apple saw the
> light eventually, and went more mainstream.
Correct. Apple's problem was simply that IBM/Motorola weren't willing
to make the investment necessary to keep PowerPC competitive with x86.
To do so would have caused losses, and they weren't making much margin
on the chips to begin with. Basically they were at the point they
didn't care if the desktop PowerPC chips succeeded or not as they only
had one customer. Embedded PowerPC is a different story. It is still
very strong, and interestingly competes in many of the same embedded
products where you see ARM--networking gear, RAID controllers, printers,
though I don't think PPC is used much in smart phones, as the power
envelop is a little too high.