Viewing page 7 of 9 pages.
Previous 1 2 3 4 5 6 7 8 9 Next
RESULTS WITH PHENOM
We recently received one of AMD's new 65W Phenom 9350E chips recently. This
is the lowest power quad-core CPU available; Intel's current lowest has a 95W
TDP. For those considering it as a low-cost, low-power quad-core alternative,
we tested it out briefly to see if was as energy efficient as claimed.
|
X2 4850e vs. Phenom 9350E
|
|
Test State
|
X2 4850e @ 2.5Ghz
|
Phenom 9350e @ 2.0Ghz
|
|
Mean CPU Use
|
Peak CPU Use
|
System Power
|
Mean CPU Use
|
Peak CPU Use
|
System Power
|
|
Idle
|
1%
|
2%
|
37W
|
1%
|
2%
|
50W
|
|
Rush Hour
|
7%
|
10%
|
~44W
|
7%
|
9%
|
~64W
|
|
Coral Reef
|
32%
|
44%
|
~53W
|
18%
|
25%
|
~72W
|
|
Flight Sim.
|
48%
|
62%
|
~67W
|
30%
|
39%
|
~77W
|
|
Drag Race
|
58%
|
72%
|
~65W
|
31%
|
40%
|
~79W
|
|
CPUBurn x2
|
100%
|
93W
|
50%
|
91W
|
|
CPUBurn x4
|
NA
|
N/A
|
100%
|
124W
|
|
CPUBurn + ATITool
|
100%
|
102W
|
100%
|
130W
|
The X2 4850e has a TDP of 45W, so the Phenom 9350e measured as specified. Power
consumption during idle and video playback was 10-14W higher, and almost the
same with two instances of CPUBurn stressing the CPU. At maximum load with all
four cores pushed hard as well as the IGP, the difference was 28W in total system
AC power or about 22W DC if you assume 80% PSU efficiency.
FAN CONTROL
For customizable fan control, SpeedFan is our application of choice. If properly
supported, it can be configured to raise/lower multiple fan speeds to designated
limits when any specified temperature threshold is breached.
|
EasyTune's Smart Fan configuration screen.
|
The screenshot above shows the correlation between SpeedFan's readings and
the temperature and fan speed sensors. Most of the readings in SpeedFan were
corroborated by Asus' PC Probe utility. Please note that though there is a "MB"
sensor, it did not seem to correlate to the chipset temperature. The MB temperature
changed very little during testing, even when we placed it on full load or put
a 80mm fan blowing over the chipset heatsink.
We found that SpeedFan could be configured to fully control both the CPU and
Chassis fan. Setting the PWM modes 1-3 in the Advanced menu from "SmartGuardian"
to "Software Controlled" enabled the speed controls. Both headers
were controllable from 0 to 100%, except in the case of a PWM fan hooked up
to the CPU fan header. A Xigmatek PWM 120mm fan connected to it could only be
controlled from between 790 to 1460 RPM approximately.
|
CPU Fan Speed (RPM)
|
|
CPU Temp.
|
Fan Profile
|
|
Silent
|
Optimal |
Perf.
|
|
30-42°C
|
820
|
830
|
1220
|
|
44°C
|
820
|
860
|
1270
|
|
46°C
|
820
|
900
|
1310
|
|
48°C
|
820
|
980
|
1340
|
|
50°C
|
840
|
1070
|
1430
|
|
52°C
|
870
|
1180
|
1430
|
|
54°C
|
950
|
1240
|
1430
|
|
56°C
|
1020
|
1300
|
1430
|
|
58°C
|
1070
|
1380
|
1430
|
|
60°C
|
1120
|
1420
|
1430
|
To test how well the automatic fan profiles worked, we connected the fan on
the CPU cooler to a variable DC fan controller set to a relatively low 6V and
proceeded to stress the processor using CPUBurn. A Xigmatek PWM 120mm fan was
connected to the CPU_FAN header and its speed was monitored and graphed using
SpeedFan. Throughout testing, the fan speed increased gradually depending on
the CPU temperature.
The range and starting and end speed was basically what differentiated the
settings. "Silent" was the most conservative with the fan speed only
having a range of 300 RPM and a high threshold temperature near 50°C. "Performance"
had a range of only 200 RPM, a high starting speed and a low threshold temperature.
"Optimal" used the fan's entire range as the CPU heat up. The Chassis
fan header produced very similar graphs — it exhibited the same basic behavior
as outlined above.
| Help support this site, buy from one of our affiliate retailers! |
|
