Viewing page 6 of 7 pages.
Previous 1 2 3 4 5 6 7 Next
Test Results - Configuration #2 (CrossFireX)
For our second test configuration we added two Radeon HD 4870's in CrossFireX.
As this increased the overall noise level, we felt it prudent to move the fan
at the top of the case to the front of the hard drive cage directly adjacent
to the graphics cards. This provides the system with some extra intake airflow
to the video card area while generating no extra noise.
|
System Measurements (CrossFireX)
|
|
State
|
Idle
|
CPU + 1 GPU Load
|
CPU + 2 GPU Load
|
|
CPU / System
Fan Speeds
|
70%* / Low
|
100% / Low
|
100% / Low
|
|
Noise
|
23 dBA
|
27~28 dBA
|
35 dBA
|
|
CPU Temp
|
38°C
|
51°C
|
58°C
|
|
SB Temp
|
61°C
|
62°C
|
69°C
|
|
HD Temp
|
39°C
|
39°C
|
39°C
|
|
GPU #1 Temp
|
79°C
|
89°C
|
99°C
|
|
GPU #1
Fan Speed
|
1100 RPM
|
2170 RPM
|
3030 RPM
|
|
GPU #2 Temp
|
79°C
|
78°C
|
89°C
|
|
GPU #2
Fan Speed
|
1150 RPM
|
1130 RPM
|
2400 RPM
|
|
*70% speed is equivalent to 8~9V
Ambient temperature: 22°C
|
The addition of two, high-end graphics cards made a noticeable
impact on system noise and temperatures. At idle, the noise level increased
by 4 dBA, and the heat radiating off the graphics cards made the CPU temperature jump
an additional 8°C. The most drastic thermal effect was on southbridge
temperature, which increased by 25°C. Not only did the cards themselves generate
more heat in the southbridge area, they also cut off airflow around the southbridge
heatsink, making it the meat in a CrossFire sandwich.
When load was applied to the system, we increased the CPU fan
speed to 100% as it did not generate any additional noise, being easily masked
by the video card fans. With the CPU and a single GPU being stressed, the CPU
heated up by only 13°C, while the southbridge and hard drive temperatures
remained virtually unchanged. The temperature of the stressed GPU went up 10°C
and the fan ramped up by almost 1100 RPM, resulting in system noise increased
by 3~4 dBA.
With both GPU's stressed, things started to get ugly. The noise
level increased by 7 dBA, and the CPU and southbridge temperatures went up by
an additional 7°C. The cooler on the 4870 occupying the top PCI-E slot was
having difficulty keeping its GPU cool. Spinning at just over 3000 RPM, it struggled
to keep to keep the GPU temperature under 100°C. No doubt this was due to
the heat radiating off its twin in the position below it, which
ran much cooler, measuring 10°C lower with its fan spinning at 2400
RPM. It should also be noted that the system power draw increased from 620W
when we began stressing the system to 636W when the temperatures had finally
stabilized. As the cards got progressively hotter, increased inefficiency resulted
in higher power consumption.
Despite the high temperatures all-around in the CrossFire setup,
the system was perfectly stable. That's not to say we would recommend such a
configuration — it just barely handled it, and noise level was unbearably
high by SPCR standards. We would recommend either using aftermarket coolers for both graphics
cards, or a case with a side fan that can blow directly over them.
Comparison
|
Antec P183 vs. Silverstone Raven
(CPU + 1 GPU Load)
|
|
State
|
Antec P183
|
Silverstone Raven
|
|
CPU Fan
Speed
|
100%
|
70%*
|
100%
|
|
Front Fan
Speed(s)
|
Low
|
70%*, 70%*
|
100%, 100%
|
|
Rear Fan
Speed
|
Low
|
N/A
|
|
Noise
|
27~28 dBA
|
27 dBA
|
29 dBA
|
|
CPU Temp
|
51°C
|
59°C
|
56°C
|
|
SB Temp
|
62°C
|
60°C
|
|
HD Temp
|
39°C
|
32°C
|
30°C
|
|
GPU #1 Temp
|
89°C
|
87°C
|
86°C
|
|
GPU #1
Fan Speed
|
2170 RPM
|
2160 RPM
|
2060 RPM
|
|
GPU #2 Temp
|
78°C
|
|
GPU #2
Fan Speed
|
1130 RPM
|
960 RPM
|
920 RPM
|
|
*70% speed is equivalent to 8~9V
Ambient temperature: 22°C
|
Compared to the Silverstone
Raven at similar noise levels, the P183 delivered better CPU temperature,
probably due to its rear exhaust fan (the Raven's 120mm rear fan was too loud
for our liking, so we removed it during testing). The Raven's two 180mm interior
fans and its bottom-to-top airflow design resulted in slightly better
GPU and southbridge temperatures as well as lower GPU fan speeds, and a significant
improvement in hard drive cooling. Though we did not test the Raven with both
GPUs fully stressed, we believe the cooling performance difference would've widened more.
If we consider just the Baseline Noise measurements, the Raven's two 180mm fans at 9V match the P183's two 120mm front and back fans at Low — both measured 17 dBA@1m. In those configurations, which is ideal for the mid/low thermal systems most silent PC enthusiasts build, the cooling of the Raven has to be superior. Its larger fans are probably moving double the volume of air of the Antec's 120mm fans.
The Antec CP-850 power supply performed well. Efficiency as measured
from the wall was similar to the Coolermaster Silent Pro we
used to power the test configuration in the Silverstone Raven. The PSU air
exhaust was warm, but not overly so, even as the
system's AC power draw increased past 600W. Look for our full review of the
CP-850 in the near future.
| Help support this site, buy the Antec Performance One P183 Black Midsize Tower from one of our affiliate retailers! |
|
