Add to Reddit
Test Results
Note : The figures stated here are all delta values (which are the difference between the CPU and ambient temperatures). They are not actual temperature values. As always, the lower the delta, the better the performance.
Results @ Idle
As you can see, the software readings are 4-5 ºC lower than the sensor readings, which are obviously way off. Bear in mind that the sensor was not placed close to the processor core.
You can see the pattern though. Even at idle, the SI-120 performed better than the XP-90. Based on the MBM5 readings, it was generally 2°C cooler, even with its fan spinning at the lowest speed.
But let's take a look at the load temperatures. They are more important anyway.
Results @ Load
Even at the lowest fan speed, the SI-120 outperformed the smaller XP-90 by at least 1°C in the MBM5 results, even though the XP-90's 92mm fan was at maximum speed. So, if you are a fan of quiet cooling, you can opt for a slower, quieter fan with the SI-120 and yet obtain better performance than the XP-90.
With the fan at maximum speed, the SI-120 was 5.5°C cooler than the XP-90 with the fan at full blast as well. That's a significant performance advantage.
We also saw a healthy reduction in PWM temperatures with the SI-120, no doubt due to the strong airflow from the 120mm fan.
Overclocking Results
And now for the overclocking test. The voltage remained set at 1.5V and ClockGen was used to raise the processor's clock speed. The Super Pi 8M test ran after every increase in clock speed until it generated an error. All tests were carried out at an ambient temperature of 22.5ºC.
Please note that in this case, higher numbers are better.
Heatsink |
Fan Speed |
Maximum Stable Frequency |
SI-120 |
2600 rpm |
3051.56 MHz |
2000 rpm |
3051.56 MHz |
|
1400 rpm |
3044.75 MHz |
|
1000 rpm |
3037.94 MHz |
|
XP-90 |
3200 rpm |
3017.50 MHz |
1800 rpm |
3017.50 MHz |
Using the SI-120 and with its fan at the higher speeds, the CPU was able to complete Super Pi 8M at 3051MHz - a full 34MHz higher than when the CPU was cooled by the XP-90. Even when the SI-120's fan was spun down to 1000 rpm, the SI-120 still had a decent advantage of 20MHz over the XP90.
In day-to-day usage, the CPU was stable for long hours of CPU-intensive D2OL crunching at 3.01GHz, whereas the magical 3GHz was not sustainable for any long periods of time.
