Philips unveils
the display of
the future, and it
looks brightIn a few years, your laptop display will be visible
from all angles, not just head-on. Your PDA will
be legible even in direct sunlight, thus minimiz-
ing the crow’s feet around your eyes from chron-
ic squinting. Your phone will last hours longer
because the display doesn’t require any backlighting. And even your lightweight,
low-profile desktop display will be able to show brilliant detail in movies and keep
up with fast-paced action games without blurring or smearing. Sounds Minority
Report futuristic, right? Well, Philips’ prototype 13-inch TV, shown below, is a liv-
ing, breathing demonstration of PolyLED’s change-everything potential.
PolyLED is Philips’ take on OLED (organic light-emitting diode) technology.
This technology dispenses with the relatively bulky semiconductors used in today’s
LCDs, which are also difficult and expensive to manufacture. Instead, PolyLED
uses a proprietary recipe of polymer materials that are sprayed onto a single layer
of glass substrate using inkjet technology. The result is a bright, self-emitting dis-
play that, in addition to using much less power, is cheaper to manufacture, and
capable of producing full-color displays less than a millimeter thick.
Philips anticipates that PolyLED televisions and monitors will debut in five
years, but we suspect market pressure from laptop, PDA, and phone manufactur-
ers—all of which are straining to eke out as much battery life as possible—could
accelerate development.GREEN TEA—BEVERAGE AND HARD DRIVE POLISH!
If you’ve ever looked inside a hard drive, you know the platters have
a mirror-like finish. That’s because they’ve been ground down and
polished to the point where surface variations are less than a single nanometer. Traditional polish-
ing fluids are a toxic slurry that must eventually be discarded somewhere (not in our backyard). But
Ventana Research has created a kinder, gentler polish using synthetic proteins and an extract from
good ole’ green tea, the favorite beverage of wise men everywhere. Because the “Ventana fluid”
binds naturally to particle debris, it’s also being considered for use in wastewater treatment.WINDOWS MOBILE 2003 SE DEBUTS
Dell’s Axim X30 is the first handheld to ship with Windows Mobile 2003 Second Edition, an update
that includes support for 640x480 (VGA) displays in landscape or portrait mode. Other tweaks
include support for dual-function cards (such as SanDisk’s Connect Plus, which combines 128MB
of storage with a Wi-Fi adapter), an update for Pocket Internet Explorer, and support for WPA (Wi-
Fi protected access) encryption. For more dirt on Windows Mobile, check out our review of the
Axim X30 next month.HD-DVD STANDARD DEFINED
The DVD Forum steering committee has approved the first physical specification for HD-DVD
(high-density DVD) read-only discs, and the spec requires that manufacturers of HD-DVD playback
devices include three video codecs: MPEG-2, MPEG-4, and the VC-9 technology used by Windows
Media Player 9 Series. Although the specification gives Microsoft’s codec a boost in credibility, it
faces competition from two competing standards—Blu-Ray, and a Chinese format called EVD. All
three are vying to establish themselves as the standard for large-capacity DVD as high-definition
television rolls out. Whether we’ll get one standard or a dreadful three-way tie remains to be seen.FUN-SIZE NEWS
Quick Start
Presenting the first image
of Philips’ prototype 13-inch
high-contrast television. We’re
assuming that’s the Discovery
Channel onscreen.Nearly everyone is ecstatic over Intel’s decision to
cancel some future single-core processors and replace
them with dual-core chips. Who doesn’t like a two-for-
one deal? However, I consider it a detour in micropro-
cessor evolution.
Since the 1970s, microprocessors have benefited from
numerous architectural innovations: pipelining, branch
prediction, speculation, hardware-assisted simultaneous
multithreading, and more. Almost every new micropro-
cessor generation has introduced a new way to run pro-
grams more efficiently—that is, to increase the number
of instructions executed per clock cycle. Integrating two
or more processor cores on a single chip does that, too.
But does that mean progress will consist largely of
slapping down more cores on a chip? Let’s hope not.
Symbolically, it would signal the end of microprocessor
evolution as we’ve known it. Pragmatically, it would
soon yield little performance improvement beyond the
usual increase in clock frequency.
Here’s why. Multicore processors try to boost perfor-
mance by executing multiple tasks in parallel. There are
two types of parallelism: instruction-level parallelism
and data-level parallelism. Most PC software has very
little instruction-level parallelism to exploit. Too many
instructions in a program need the results of previous
instructions before they can execute, so they must wait
until those previous instructions finish computing their
results. Imagine a factory worker on an assembly line
trying to mount the car’s wheel before the axle has been
installed, and you’ll grasp the problem. Even if your CPU
chip had 100 processor cores, it wouldn’t run most of
your productivity apps much faster than they run now.
Data-level parallelism is easier to exploit, at
least with programs that have lots of graphics, audio,
video, or other large data sets. A great example of
data-level parallelism is the SETI@Home (Search for
Extraterrestrial Intelligence) project. It uses the Internet
to distribute chunks of digitized radio noise received
from outer space to about 500,000 computers around
the world. Each PC independently analyzes its chunk of
data, looking for patterns that might indicate an intel-
ligent radio signal. A virtually unlimited number of pro-
cessors can work in parallel on this task, because one
chunk of data doesn’t necessarily depend on a result
from another chunk of data.
That’s why graphics-rich games, image editors, sound
editors, and other data-intensive apps stand to benefit
from multicore processors. Most other kinds of programs
will hardly benefit at all. So the future progress of com-
puting still depends on the evolution of the processor
itself, not simply on replicating many processors on a
single chip.The CPU’s
Evolutionary
Detour
14 MAXIMUMPC AUGUST 2004
Tom Halfhill was formerly a senior editor for Byte magazine and
now an analyst for Microprocessor Report.FAST FORWARD BY^ TOM R. HALFHILL Presenting
PolyLED
Green tea: Good for you,
good for your hard drive.