Is this the siren song for the x86 architecture
and its great bastion, Intel? It's hard to say for sure, but Microsoft's
official announcement that it was supporting a more
efficient rival architecture -- ARM -- certainly was met
with little joy in Santa Clara.
At the show Microsoft CEO Steve Ballmer showed off
a series of development systems running a "next generation version of
Windows", which supported ARM. Microsoft layered Windows 7's
graphical user interface on top of new OS to show just how smoothly an ARM
powered Windows system could run.
In total Microsoft showed off three different ARM
development systems, with a system-on-a-chip design from Qualcomm (SnapDragon),
Texas Instruments (OMAP), and NVIDIA (Tegra 2). Mr. Ballmer did not
officially announce when we might expect to see these Windows ARM systems, but
it might be sooner than you think.
Microsoft has grown increasingly impatient with
long-time partner Intel, who manufacturers somewhere between 80 to 90 percent
of the world's computer CPUs. Intel was being badly beaten in the fight
for smartphone and tablet dominance -- or more aptly it never
showed up, because it knew it was a fight that it couldn't win.
Microsoft had already long since gone with ARM
processors in the ultra-power dependent smartphone industry. But in the
tablet sector it sat by
and watched in pain as Apple and Google unloaded ARM based designs
by the millions. There were no Windows 7 tablets because Intel was unable
to provide it hardware.
Unwilling to see its hopes anchored to what may be
a sinking ship, Microsoft made the tough decision to jump onboard the ARM
train, a serious vote of no-confidence for x86. The message seemed clear
-- Intel's promises
of Atom-based Windows 7 tablets were welcome, but Microsoft sure wasn't waiting around
for their release.
As ARM suppliers gains momentum they are hungrily
eyeing the netbook, notebook, and PC markets. Already we're seeing
dual-core ARM CPUs show up in smartphones, and there's talk of eight-core
ARM CPUs clocked as high as 2 GHz being delivered within a
generation or two. So is Intel's CPU (and to a lesser extent those of
AMD) destined for a slow ride into the sunset, replaced by NVIDIA, Qualcomm,
Texas Instruments, and Samsung chips?
It's harder to say. ARM's great hope is that
it can parlay its sizable lead in power efficiency over x86 into market
dominance. ARM features a reduced-instruction set, versus Intel's
cluttered instruction catalog. And it has more integers registers, which
eliminates the expensive process of renaming registers. The net result of
both of these architectural differences is that ARM can perform the same
computation using less power.
And yet Intel could still pull out a
victory. As circuits have shrunk, leakage of current from the capacitors
inside transistors has become a major issue. In today's generation of ARM
and x86 CPUs, leakage can account for as much as 40 percent of the power
consumption of a chip. As leakage becomes more important, process
technologies may become more important, while subtle architectural advantages
become more trivial.
Thus if Intel can hold on, it may stand a shot,
thanks to its tireless advances in the field of process technology, which include "high-K dielectrics" -- special
capacitor materials that combat leakage.
On the other hand, developing processes is an
expensive business, and if ARM begins a successful campaign into the personal
computing market, it may starve Intel of the capital it needs to survive.
One thing is for sure -- for now consumers have
compelling cause to buy ARM OS tablets, netbooks, and notebooks, a cause
Microsoft has recognized and addressed. Intel can only hope to weather
quote: You don't see people trying to build an ARM-based desktop