reviews TESTED. REVIEWED. VERDICTIZED
T
his month, CoolIT’s Boreas rocks us
with a King Kong-like experience by
brute-forcing our CPU to new levels
of cooling. Gigabyte’s rocket cooler, mean-
while, attaches to our test bed with the
same strength the Space Shuttle is affi xed
to its launch pad. Will these CPU coolers
be the death of the fabled Maximum PC
testing apparatus?
—DAVID MURPHY
COOLIT BOREAS
There comes a time in every young PC
builder’s life when he seriously considers
outlandish ideas for modifying and cool-
ing his smokin’ new gaming rig. But you
don’t need to mod your PC into a refrig-
erator to reach subzero temperatures, not
if you have CoolIT’s latest 12 TEC cooler,
the Boreas.
CoolIT’s Boreas could likely fi t a
stock CPU cooler inside its sprawling
gut, a 4-inch-long heatsink with a 12cm
fan on each end. Two mounting brackets
are preattached to the device, and thank
heavens! Few cases have enough room for
this chubbo. The included MTEC control
box connects the Boreas’s automation
software to the device, although it uses
an entire USB header with absolutely no
pass-through.
That’s not too bad a sacrifi ce given the
payoff. On its quietest setting, the Boreas
beats out the loud Zalman CNPS9700
air cooler in both our idle and burn tests.
Cranking the Boreas to the max drops our
CPU temperature to 11 C, but created
dangerous levels of condensation. It’s
the Boreas’s Achilles’ heel: Cool too
much and you could destroy your rig.
The included software and control-
ler mechanism minimize this issue by
letting you pick a temperature for the
Boreas to shoot for—a feature sorely
missing in previous CoolIT coolers—
but fi nding the perfect setting requires
a little experimentation.
GIGABYTE 3D ROCKET II
Let nobody say that Gigabyte didn’t break
the mold with its 3D Rocket II heatsink/fan
combination. As the name alludes, the device
resembles a rocket ship sitting atop a launch
pad. It’s about as well strapped in, too; we
applaud the 3D Rocket II for its efforts to sail
amongst the heavenly stars of CPU coolers,
but its installation process keeps the device
strapped fi rmly to the ground.
We fi rst tried to mount the cooler in our
old FX-60 test bed, a relatively painless
procedure given how nicely the cooler’s
single retention bar snaps
into the motherboard’s
mounting mechanism.
However, the pressure the
cooler placed on the CPU
caused our motherboard
not to post. The damage,
though, wasn’t permanent.
Everything worked perfectly
after we swapped out the 3D Rocket II for
a stock AMD cooler.
Installing the cooler on our new
Intel Q6700-based test bed forced us
to remove the motherboard to install
Gigabyte’s mounting bracket. We were
worried about possibly breaking our CPU
by having to push down on the cooler’s
two retention bars, but the machine
posted and gave us results comparable to
those of our current air-cooling champion.
The tall cooler barely fi ts in most cases,
forcing you to remove the optional air duct
attached to its fan. But that doesn’t trouble
us nearly as much as the possibility that the
3D Rocket II could grind our processors into
dust. Anxiety is the last thing we want in an
aftermarket CPU cooler.
Cooling Cavalcade
Chilly temperatures can come with a price
The Gigabyte 3D Rocket II comes
with more accessories than we’ve
ever seen in a cooler.
$45, http://www.gigabyte-usa.com
GIGABYTE 3D ROCKET II
MAJOR MAJOR
Comes with decorative
UV-reactive fluorescent rings
and a fan speed controller.
MAJOR TOM^7
Pushes on the CPU a bit too
much; design makes for a
tricky installation.
$450, http://www.coolitsystems.com
COOLIT BOREAS
KRATOS
You will not find a more
powerful cooler.
ARES^9
Ear-splittingly loud when
cranked up, uses an entire USB
header, tricky condensation.
For its bulk, the CoolIT Boreas is surprisingly quiet
when its fans are on their lowest setting.
76 MAXIMUMPC | APR 08 | http://www.maximumpc.com
23.8 11.0 31.5 37.0 43.8
44.8 24.3 48.5 46.0 67.5
COOLIT BOREAS COOLIT BOREAS GIGABYTE 3D ZALMAN STOCK
(LOW) (HIGH) ROCKET II CNPS9700 COOLER
BENCHMARKS
IDLE (C)
100% LOAD (C)
Best scores are bolded. Idle temperatures were measured after an hour of inactivity; load temperatures were measured after
an hour’s worth of CPU Burn-In (four instances). Test system consists of a stock-clock Q6700 processor on an EVGA 680i
motherboard with an Nvidia 8800 GTX graphics card.