MAXIMUM
PC’
s
BUILD A
HOME
THEATER PC!
We’ll show you how to build a
home theater rig so powerful—yet
quiet—you’ll be able to charge
admission to your living room!HOME
THEATER
PC CASE
ROUNDUP!
The journey to the ultimate home
theater rig begins with your case
selection. Let us do the legwork.
All the top HTPC cases will be
reviewed and verdictized!TECH
TRAGEDIES!
AKA the “What were they think-
ing?” awards, we’ll chronicle the
top 10 tech disasters, debacles,
and atrocities—of all time!NEXT
IN
SEPTEMBER 2006 MAXIMUMPC
BACKSTAGE AT
OZZFEST
LETTERS POLICY: MAXIMUM PC invites your thoughts and comments. Send them to
[email protected]. Please include your full name, town, and telephone number, and limit
your letter to 300 words. Letters may be edited for space and clarity. Due to the vast amount of
e-mail we receive, we cannot personally respond to each letter.CONTRIBUTING EDITOR TOM HALFHILL
RESPONDS: Sorry, Jack, I don’t buy your argu-
ments for net neutrality. The exclusionary
behavior you describe would almost certainly
be ruled anti-competitive and doesn’t happen in
the marketplace today. For instance, FedEx owns
Kinko’s, and UPS owns Mail Boxes Etc., but both
companies still deliver packages to each other’s
locations and customers in a timely fashion.
What would happen if FedEx or UPS stopped
delivering to recipients that have business rela-
tionships with competitors?
Your second argument falls apart because
your scenario already happens. When a packet
times out, for whatever reason, do you ever get an
explanation? Right now, some packets get higher
priority than yours, except the senders aren’t pay-
ing for it. Suppose your email arrives late or not
at all because Victoria’s Secret staged an online
event that hogged all of your provider’s network
bandwidth. Victoria’s Secret didn’t pay extra for
that bandwidth. What is your recourse now?
Your third argument also describes today’s
Internet. The phone company could give DSL to
everyone, not just those who pay for it. Isn’t that
an “artificial performance constraint”? Dial-
up isn’t slower just because DSL is faster. My
monthly DSL fee pays for new equipment that
adds bandwidth. The same will happen when
Amazon, Google, Microsoft, Yahoo, Victoria’s
Secret, and other bandwidth hogs start paying
their fair share for the resources they consume.
Should net neutrality come to pass, I predict
that advocates will plead for tiered service in a
few years. A single downloaded feature film may
use more bandwidth than all the emails and text
messages you will send in your lifetime. As the
Internet chokes on multimedia streams requiring
isochronous delivery, you will see the wisdom of
charging users higher rates for heavier freight.
$2,000 GRAPHICS CARD?
I’ve been looking at quad-SLI systems, because
I’m willing to pay anything to get the best 3D
performance. However, I recently found an interest-
ing-looking card on Pricewatch.com called the
Wildcat Realizm 800. The specs look crazy (256GB
virtual memory, dual GPUs, 3840x2400 resolution,
and 640MB of GDDR3 memory). Is this board worth
$2,000? Is it better than quad-SLI?
—MIKE BECK
EXECUTIVE EDITOR MICHAEL BROWN RESPONDS:
It depends on what you want to use your video-
card for. The Wildcat Realizm 800 is based on
a 3DLabs GPU and is designed for accelerating
workstation applications, such as AutoCAD and
Maya. It should be compared to other worksta-
tion cards, such as ones based on nVidia’s
Quadro and ATI’s FireGL GPUs. Comparing work-
station cards to a GeForce quad SLI is compar-
ing apples to oranges.
None of the workstation cards are well suit-
ed to games; some don’t even include DirectX
drivers. But if you’re more interested in design-
ing than playing, they’re all worth a look.PHYSICS FOLLOW-UP
While I’m disappointed to hear about the lackluster
visual performance of the new PhysX card reviewed
in the July issue, I have one question regarding
the review. Supposedly, one of the benefits of the
physics accelerators is that most or all of the phys-
ics processing work is offloaded from the CPU and
handled by the physisc processor, in turn leading to
faster performance. Did you test for a performance
change? If so, what were the results? If not, would it
be possible to include those results in future physics
accelerator reviews?
—Matthew SniderEXECUTIVE EDITOR MICHAEL BROWN RESPONDS:
It’s not easy to evaluate a piece of hardware
when it establishes an entirely new market seg-
ment, and this difficulty is compounded when
there’s very little retail software designed to
take advantage of it. It’s like trying to measure
the performance of new tires without the benefit
of a car and a test track. So it wouldn’t be fair to
characterize the PhysX chip as having delivered
“lackluster visual performance.” With so few
games available for testing, it’s impossible to
know whether the problem is inherently related
to the processor, or if the software just did a
poor job of leveraging the processor’s abilities.
When a real game running on the PhysX chip
looks as good as the Cell Factor demo, we might
have a different opinion of the chip.
As for offloading physics processing from
the CPU, there’s so little advanced physics in
Ghost Recon: Advanced Warfighter that install-
ing the PhysX card delivered an inconsequen-
tial boost in frame rate, in fact, we experienced
the opposite. Because the graphics card is
rendering many more objects, we actually
experienced a slight, but noticeable slowdown
in GRAW. And because Cell Factor runs only
when the card is present, there’s no way for us
to make a comparison.OCTOBER
ISSUE
COMING
MONTH