A fanless, silent version of the Asus 8600GT graphics card turned up in our lab recently. This is identical electronically to another Asus card we reviewed recently, except it has 512mb of RAM. Is the passive cooling solution good enough for a hot, low airflow, silent system?
November 11, 2007 by Lawrence
Lee
| Product | Asus EN8600GT Silent/HTDP/512M PCI-E Video Card |
| Manufacturer | ASUSTeK |
| Market Price | US$145 |
Recently we reviewed the Asus
EN8600GT OC GEAR graphics card and came away with slightly mixed feelings
about the cooling solution used. The fan employed was unnecessarily loud in
our opinion, as we found that lowering the fan speed did not compromise the
GPU temperature when stressed. Today we look at the Asus EN8600GT Silent, effectively
the same card only this time, passively-cooled. In fact, Asus claims it runs
“7°C cooler than generic boards” — a bold statement.
|
|
|
|
| Asus EN8600GT OC GEAR/HTDP/256M: Specifications (from the product web page) | |
| Graphics Engine | GeForce 8600GT |
| Video Memory | 512MB DDR3 |
| Engine Clock | 540MHz |
| Memory Clock | 1.4GHz (700MHz DDR3) |
| Memory Interface | 128-bit |
| CRT Max Resolution | 2048×1536 |
| Bus Standard | PCI Express X16 |
| VGA Output | YES, via DVI to VGA Adapter |
| HDTV Output (YPbPr) | YES, via HDTV Out cable |
| TV Output | YES, via S-Video to Composite |
| DVI Output | DVI-I |
| Dual DVI Output | YES |
| Adapter/Cable Bundled | DVI to VGA adapter HDTV-out cable |
| Software Bundled | ASUS Utilities & Driver |
THE CARD AND HEATSINK
This 8600GT is essentially identical electronically to the OC Gear version
we reviewed previously, except the amount of video memory has been doubled and
the heatsink replaced with a passive solution. The core and memory run at the
same frequencies.
|
|
The heatsink’s fins are very thick and spaced widely apart. A shroud helps
focus air current over the heatsink. It’s doubtful this will be useful without
appreciable airflow. It’s notable that unlike some other passively cooled graphics
cards from Asus, (such as the EN7600GT
SILENT, or the EN8600GTS
SILENT) the heatsink does not extend to the backside of the board. Presumably,
this has to do with the actual cooling requirements of the different GPUs on
these graphics cards.
|
|
|
|
The bulk of the cooler appears to be comprised of aluminum but a different
story emerges when viewed from the side. The actual GPU core contacts a thin
copper base plate. Just above it, a single solitary copper heatpipe curves out
the opposite side and transfers heat to the rest of the heatsink. A small section
of the heatpipe is visible on the top side of the cooler — the shroud obscures
the rest from view.
|
|
The fin layout of the heatsink is puzzling. Instead of a symmetrical array,
the fins are broken into sections. The middle portion is curved and the section
at the top of the PCB is arranged at a steep angle.
|
|
From the design it appears that Asus intends warm air to exit via the vent
and out of the case, but we’re not sure exactly how this can be accomplished
without any direct airflow. Ultimately whether air enters or exits this grill
will be determined by the internal case pressure. The “dorsal” fins
(on the left side of the above picture) are almost perpendicular to the center
fins. It seems all they do is get in the way of any air that may be flowing
down the length of the PCB. A more uniform design with a higher quantity of
thinner fins to maximize the amount of surface area exposed would serve better
in a low airflow environment.
|
|
TEST METHODOLOGY
|
|
Our first test procedure is an in-system test designed to determine the card’s
power consumption, and whether the card is capable of being adequately cooled
in a low-noise system. By adequately cooled, we mean cooled well
enough so that no misbehavior related to thermal overload is exhibited. Thermal
misbehavior in a graphics card can show up in a variety of ways, including…
- Sudden system shutdown or reboot without warning.
- Jaggies and other visual artifacts on the screen.
- Motion slowing and/or screen freezing.
Any of these misbehaviors are annoying at best and dangerous at worst – dangerous
to the health and lifespan of the graphics card, and sometimes to the system
OS.
Test Platform
- Intel
Pentium D 930 Presler core processor. Official TDP of 95W. - AOpen
i945Ga-PHS motherboard – Intel i945Ga Chipset; built-in VGA. - CoolerMaster Hyper 48
heatsink, cooled by a Nexus 92mm fan undervolted to 7V. - Corsair
DDR2 RAM, 1024 MB - Seagate Momentus 5400.2
120 GB, 2-platter drive, suspended in a NoVibes
III just inside the front intake vent. - Antec Neo HE 430
ATX12V 2.01 compliant power supply, with a custom-built fresh air duct to
ensure that the internal fan did not ramp up during testing. - Modified case from Cool
Cases, outlined in detail below. - Nexus 120mm
fan controlled by a variable voltage fan controller.
Measurement and Analysis Tools
- CPUBurn
processor stress software - ATITool
v0.26
artifact scanner to stress the GPU. - RTHDRIBL
as an alternate tool for stressing the GPU. - SpeedFan
version 4.33 to show CPU and GPU temperature. - Seasonic
Power Angel AC power meter, used to measure the power consumption
of the system. - A custom-built variable DC power supply to power the system fan.
System airflow is quite good, allowing the CPU and system fans to run at close
to inaudible speeds without compromising system cooling. The intake is about
the size of a 120mm fan. The only restriction is an air filter. A much more
restrictive cover for the filter was removed because it impeded the airflow
too much.
|
|
|
|
There is only one point of exhaust: The Nexus 120mm case fan which will be
run at a number of different speeds. The 80mm fan in the Neo HE power supply
was taken out of the picture by using a custom-built duct to ensure that the
fan never ramped up. The amount of airflow through the system can be controlled
by adjusting the speed of the case fan, thereby giving us a way of controlling
how difficult the thermal environment inside the case is.
|
|
|
|
The airflow in our test rig is typical of an ATX case. Air flows in through
the intake near the bottom of the front panel, and is pulled up to the top rear
corner. Most of this air will bypass the expansion cards altogether, but a small
amount will be pulled across the rear of the card as it is pulled towards the
CPU heatsink and the case fan. All of the air will exit the case via the exhaust
fan.
|
|
Testing was conducted initially with the system fan at 12V with
the system idle, and running two instances of CPUBurn. Then ATITool’s artifact
scanner was run in conjunction with two instances of CPUBurn to generate as much
heat and power draw as possible from both the CPU and GPU. During this portion
of testing the system fan voltage was slowly turned down to see what thermal
consequences arose from limiting the amount of system airflow provided.
SpeedFan and was used to measure CPU and GPU temperature, and the results were
recorded after the GPU temperature stabilized. After the final, most demanding
test with the system fan running at 7V, ATITool was left running for 30 minutes
and then replaced by RTHDRIBL for 30 minutes after that. Throughout this step
we watched the screen for visual artifacts that might indicate overheating.
If either ATITool or RTHDRIBL detected artifacts in this minimum airflow and
maximum heat test situation, the card would be retested for artifacts with the
system fan at 9V and 12V to determine whether system airflow was a contributing
factor to failure.
The card is determined to have passed our testing if it managed to survive
the duration of testing without ATITool or RTHDRIBL detecting any artifacts.
AC system power consumption was measured from the outlet using a Seasonic
Power Angel and actual DC power use was estimated/extrapolated using the efficiency
figures deduced from our Antec
Neo HE 430 review: 67% at 62W, 76% at 85W, 76% at 119W, 79% at 190W
(all wattages are AC).
During testing the ambient temperature was 21°C and the noise level was
approximately 18 dBA.
TEST RESULTS
Before getting to the actual test results, we first offer a baseline for comparison:
results from the test system using the integrated video on the motherboard at
the same ambient temperature. Note, neither ATITool artifact scanner or RTHDRIBL
would run on the integrated graphics chip. We used ATITool’s 3D View instead
as we know from past experience that it generates a similar amount of power
consumption but without the ability to detect artifacts.
| VGA Test Bed: Baseline Results (no external VGA card) | |||
| System State | CPU Temp | System Power | |
| AC input | DC output (Est.) | ||
| Idle | 24°C | 80W | 59W |
| CPUBurn | 50°C | 151W | 117W |
| CPUBurn + ATITool | 50°C | 161W | 125W |
| System fan @ 7V, 21 dBA @ 1m | |||
| VGA Test Bed: Asus EN8600GT Silent | |||||
| System Load | System Fan Speed | GPU Temp | CPU | System Power | |
| AC input | DC output (Est.) | ||||
| Idle | 12V | 55°C | 23°C | 100W | 76W |
| CPUBurn | 12V | 56°C | 48°C | 171W | 134W |
| CPUBurn + ATITool | 12V | 91°C | 44°C | 188W | 148W |
| CPUBurn + ATITool | 9V | 98°C | 45°C | ||
| CPUBurn + ATITool | 7V | 103°C | 50°C | ||
With the system fan running at the default 12V, the GPU reached 91°C,
which isn’t too bad for a passive heatsink. When we dropped the fan to 9V,
the temperature rose sharply by 7°C. The fan became inaudible in our test
system at 7V and as if on cue, the GPU broke the 100°C mark. It’s clear
that there’s a high price to pay for silence, and system airflow is incredibly
important for a passively cooled graphics card. The CPU temperature also rose
by 5°C during this final test — the heat rising from the video card
obviously affected the processor’s temperature.The good news is that throughout testing, no instability was exhibited and
no artifacts detected by either ATITool or visual inspection of RTHDRIBL.
Our test system is very tough thermally as the only component providing system
cooling airflow is a Nexus 120mm exhaust fan, which does not push a lot of
air at any speed. Still, the EN8600GT Silent passed this portion of testing,
though at higher temperatures than we’d like. Our tests show that a little
bit of airflow goes a long way, so if your system has an extra fan, you likely
won’t experience as high temperatures as we did.
| Graphics Card Thermal Comparison: Asus EN8600GT Silent vs. EN8600GT OC Gear | |||
| Graphics Card | SPL @ 1m | GPU | CPU |
| EN8600GT Silent | 21 dBA | 103°C | 50°C |
| EN8600GT OC Gear (fan @ minimum speed) | 22 dBA | 70°C | 45°C |
| System fan @ 7V | |||
With a much larger heatsink and no direct airflow, the Silent edition of
the EN8600GT ran 30°C hotter than the OC Gear version with only a slightly
lower noise level.
Power
A rough idea of the power consumption of the card can
be judged by comparing the total system power consumption with and without
the card installed. Our results were derived at by assuming:1. The power consumption of the graphics card at idle
was equal to the difference in power demand between the two systems when both
were running CPUBurn.2. The power consumption of the graphics card under load was equal to the
difference between the system with the card running CPUBurn and ATITool simultaneously,
and the baseline system running CPUBurn only. This ensured that any load on
the CPU from ATITool did not skew the results, since the CPU was running
at full load in both cases.
| Graphics Card Power Consumption Comparison: Asus EN8600GT Silent vs. EN8600GT OC Gear | ||||
| GPU State | EN8600GT Silent | EN8600GT OC Gear | ||
| Increase (AC) | DC (Est.) | Increase (AC) | DC (Est.) | |
| Idle | +20W | 15W | +23W | 18W |
| Load | +37W | 29W | +45W | 36W |
In an odd twist, we found that the Silent edition of Asus’ 8600GT consumed
significantly less power than the OC Gear even though it sports twice as much
memory. We were only expecting a difference of a watt or two because there
was no fan being powered by the card this time around. The drivers used were
the same, no additional software had been installed and the OC gear unit was
not hooked up during its thermal and power testing. It’s fantastic that it
consumes 7W less at full load than the fan cooled version. This is probably
explained as improvements in GPU core efficiency in later production runs,
and simple sample variance… but we simply do not have enough experience
with big batches of the same GPU core types to say for sure.
VIDEO PLAYBACK TEST
Our second test procedure is designed to determine the card’s proficiency at
playing back high definition videos encoded with the popular H.264 and the up-and-coming
VC-1 codecs. We used the same test platform as our thermal and power test with
the Intel Pentium D 930 Presler processor. It’s a fairly low-end dual core CPU
by modern standards, so the difference in the amount of assistance it lent to
the GPU between the different test videos would be more easily distinguishable.
The clips were played with Windows Media Player 11 and a CPU usage graph was
created by the Windows Task Manger for analysis to determine the mean and average
CPU use. The higher the CPU usage, the lower the video card’s decoding ability.
If CPU usage reached extremely high levels and the video skipped or froze, we
concluded the video card failed to adequately decompress the clip. System power
consumption was also recorded.
For complete details of the video clips used, please see page
5 of the Asus EN8600GT OC GEAR graphics card review.
VIDEO PLAYBACK TEST RESULTS
| CPU Usage & Power Consumption | ||||
| Video Clip | Mean CPU Usage | System Power Consumption (AC) | ||
| Core 0 | Core 1 | Average | ||
| 720p H.264 | 27% | 6% | 16.5% | ~117W |
| 1080p H.264 | 44% | 28% | 36.0% | ~131W |
| WMV3 VC-1 | 39% | 22% | 30.5% | ~131W |
| WVC1 VC-1 | 56% | 46% | 51.0% | ~144W |
| Graphics Card Video Playback Comparison: Asus EN8600GT Silent vs. EN8600GT OC Gear | ||||
| Video Clip | EN8600GT Silent | EN8600GT OC Gear | ||
| Average CPU Usage | System AC Power | Average CPU Usage | System AC Power | |
| 720p H.264 | 16.5% | ~117W | 16.0% | ~120W |
| 1080p H.264 | 36.0% | ~131W | 36.5% | ~132W |
| WMV3 VC-1 | 30.5% | ~131W | 30.5% | ~133W |
| WVC1 VC-1 | 51.0% | ~144W | 53.0% | ~144W |
Playback results were excellent and consistent with the EN8600GT OC Gear,
as expected. This test was more of a formality than anything else. Interestingly,
the Silent model demanded slightly less power, but the difference was not
nearly as big as during the maximum load testing.
NOISE RECORDINGS IN MP3 FORMAT
These recording starts with 4~10 seconds of “silence” to let you
hear the ambient sound of the room, followed by 10 seconds of the test system
noise (the Asus EN8600GT Silent does not produce any noise). The recording of
the Asus EN8600GT OC Gear video card has its fan at three settings: Minimum,
default, and maximum fan speeds. There’s a few seconds of “silence”
inserted between each 10 second stretch of noise to help you remember the reference
ambient.
- VGA
Test System with no Video Card, system fan at 7V, @ 1m (21 dBA) This
is exactly the same noise obtained with the EN8600GT Silent, since this card
produces no noise. - System
with Asus EN8600GT OC Gear, at minimum, default and maximum fan speed @1m
(22, 24 and 27 dBA)
| HOW TO LISTEN & COMPARE These The The one foot recording is More details about how we make these recordings can be found in our short article: Audio Recording Methods Revised. |
CONCLUSIONS
We cannot confirm nor deny the claim made by Asus that the EN8600GT
Silent is “7°C cooler than generic boards.” The use of the word
“generic” really makes this statement ambiguous. Does it mean other
passively cooled 8600GTs or are they are referring to 8600GTs that use the nVidia
reference cooler? If it’s the former, it’s quite possible. If it’s the latter,
we seriously doubt it. It is unlikely that an actively-cooled 8600GT would ever
hit over 100°C in our test system. If the EN8600GT OC Gear is any indication,
a fan at any speed would improve cooling by leaps and bounds.
For a graphics card that occupies two slots, we were expecting
better thermal results — Asus’ heatsink design could use some improvement.
We simply don’t have information to know what is a safe long term operating
temperature limit, but the >100°C reached at the lowest system fan setting
seems too high.
You will want to design more directed airflow for the graphics
card (or more overall case airflow) than we have in our test setup when the
Nexus case fan is running at 7V; this is about as little airflow as any hot
system should ever have. It’s likely that the Asus folks did not intend this
graphics card to be used in a case with so little airflow.
The cooler worked well enough even under such tough thermal conditions
to keep the card operating without any anomalies, so despite the high GPU temperatures,
we give this card a cautious recommendation along with a caveat to ensure good
case cooling. As for the other criteria we consider important — silence,
power consumption, and video playback ability — the Asus EN8600GT Silent
is a slam dunk.
Many thanks to ASUSTeK
for the sample of the Asus EN8600GT Silent/HTDP/512M
* * *
SPCR Articles of Related Interest:
Asus EN8600GT OC GEAR graphics
card
Sytrin KuFormula VF1 Plus
graphics card cooler
Fanless PCIe Graphics Cards
from Asus and Aopen
Gigabyte GV-N66256DP Fanless AGP video card
* * *
