MaximumPC 2003 12

(Dariusz) #1

46 MAXIMUMPC DECEMBER 2003


You can call it a cosmic silicon
convergence. Such is the signifi-
cance of the three biggest names
in personal computing all releas-
ing new desktop CPUs within
weeks of each other. We wouldn’t
even begin to posit what this says
about the vitality of the comput-
ing industry, but we can tell you
that this once-in-a-millennium
event makes for a damn good
benchmarking showdown.
Apple came to market first
with its new G5 platform. The G5
is based on IBM’s PowerPC 970
chip and offers more memory
bandwidth than Mac users have
ever previously enjoyed. Next
up was AMD’s ground-breaking
Athlon 64 FX, which introduces
64-bit computing to the PC desk-
top. And a few days after AMD
announced the 64 FX, Intel com-
pleted the triangle with its Pentium
4 Extreme Edition, which offers
more than twice the cache of its
AMD competitor.
Thus the battle lines were
drawn. This month we offer

you definitive real-world road-
testing of each CPU, using cross-
platform applications that you
can run at home.
Now, we’ll be the first to
concede that not all versions of
the same application are afforded
equal attention by their developers.
The PC version of any given app
might be coded to take advantage
of the latest PC platform technolo-
gies, but the Mac version of the
very same app might not take
full advantage of the latest Mac
technologies—and vice versa.
Sometimes, software simply needs
to “catch up” with new hardware
platforms in order to take full
advantage of new instructions,
multithreading, and so on. That
said, one can’t really judge a CPU
platform in isolation. The software
that runs on the platform must
be considered as well, and if the
software just plain isn’t ready, then
that’s a factor that must be consid-
ered in any consumer’s final hard-
ware decision.
To help choose the right
cross-platform benchmarks, we
ran a list of proposed tests past
the folks at MacAddict , one of our
sister magazines. For the most
part, they blessed our selection
of cross-platform tests. They also
had a hand in creating one of our
Photoshop test scripts, and sug-
gested we use the cross-platform

app Mathematica because of its
popularity with Mac users.
We also obliged Apple
Computers by running some of
the very tests that the company
references in its own benchmark-
ing documents. One is MacBibble ,
an app that converts RAW image
files into JPEGs (the PC version of
this app is simply called Bibble ).
We also fulfilled Apple’s request
that we use the Mac application
Compressor to benchmark DV-to-
MPEG2 encoding times. As there
is no PC version of Compressor ,
we used a similar app called
ProCoder to test encoding on the
PC side. Obviously, Compressor
and ProCoder are entirely different
apps, so comparing performance
with these packages is sketchy.
They both shoot out MPEG-2 files,
but it’s debatable whether the
visual quality of each resulting file
is identical.
Conducting cross-plat-
form comparison tests is
more difficult than ever before.
Nonetheless, we’re confident our
showdown will reveal whether
the G5 is truly the “fastest PC on
the planet” (not just on paper,
but in an actual working environ-
ment), and whether the Pentium
4 Extreme Edition grabs the
performance crown back from
AMD’s Athlon 64 FX.
Let the battle begin!

It’s Intel vs. AMD


vs. Apple in a


bloody battle for


CPU dominance!


Three enter our


benchmarking cage


match, but only one


seals victory!


BY GORDON MAH UNG
Free download pdf