Re: Advice on system purchase
On 10/29/2012 6:08 PM, Celejar wrote:
> Interesting. Google shows that there was a thread on /. a year ago
> about the question of ARM on the desktop, but a quick skim shows no
> obviously compelling reason why it won't ever happen. Thoughts?
There a dozens of reasons. First and foremost, ARM sells millions of
chips into the portable device and embedded markets. If they could
break into the desktop PC business their unit shipments would be tiny,
and thus not worth the platform investment.
The second big reason is that neither Microsoft nor ISVs will profit
from a non x86 CPU architecture entering the desktop space. Supporting
ARM would simply cost them money. So there's no incentive to support
ARM, thus it's dead before it gets started on the desktop.
Third, "Wintel" is more than just nickname. It's already not in MS'
interest to support ARM, and it's definitely not in Intel's interest, so
when Intel says "don't do it", it's dead.
At this point in time, and in the foreseeable, the only way to crack
into the desktop market is with a new x86 chip that has sufficiently
compelling advantages over both Intel and AMD. And since one must have
a license from Intel to do so, that ain't gonna happen.
If Linux takes over 50% of the desktop in the future, an ARM desktop
system becomes much more likely. But Linux isn't going to own 50%, let
alone 20%, of the desktop any time soon.
Note that PREP/CHIRP failed as well as the AlphaPC initiative. The
desktop CPU market is a tough nut to crack. Of Intel's x86 competition
Cyrix was swallowed by VIA whose processors don't sell in the Western
hemisphere, Transmeta fell off the map, AMD swallowed NexGen and is
barely hanging on in the x86 desktop/server CPU space.
In summary, ARM won't enter the desktop CPU space because the potential
profit, if any, is too small, the likelihood of huge losses too high,
and the barriers to entry are too high.