The Test
With each cooler we test, we are able to add another set of data points on our cooling performance charts. It is very convenient to have a system with a constant lineup of components to test these coolers on, as individual results from one test at FPSLabs can be compared to the corresponding results of another.



**Please note that all of our testing was done on a quad-core processor with a 120W TDP (actual value is more likely around 140-150W). Quad-core processors (specifically, Kentsfield-based chips from Intel Corp) produce quite literally twice the heat as their dual-core brethren, and as such require the use of different cooling philosophies. There are different types of heat transfer. Coolers reviewed on sites like Anandtech will receive final results different than those shown on our site because Anandtech uses a dual-core processor for all of their thermal tests (as of writing, September 07). Typically, coolers that score very well on Anandtech are capable of moving (smaller amounts of) heat very quickly and also getting rid of it quickly through convection, thus producing lower temperatures. Such coolers might not necessarily perform well at FPSLabs because the processor we use to test produces far more heat and is a much more stressful task for the cooler. Thermal solutions that perform well in our tests might not move heat as quickly as top-performing coolers on Anandtech, but far more heat is being moved and removed, just at higher overall temperatures.

This cooler was extremely impressive in our thermal testing. Not only did it trump even the Monsoon II Active TEC Cooling solution from Vigor Gaming at idle conditions, it also beat that augmented cooler, albeit by a small margin, under load. This is a tremendous feat for any pure air cooler, especially one with a 92mm fan. Not once during testing did the cooler - at least the parts of it we were able to reach - get warm to the touch. The exhaust coming from the back of the cooler was nowhere near as warm as with the Monsoon II, which adds an extra 50W of heat to the mix from its integrated TEC plate. Also, please remember that the testing methodology we use calls for reading temperatures from the on-chip diode that reports the Tjunction temperature of the processor. As this diode is part of the actual CPU die itself, the temperatures it reports, we feel, are very important. However, the temperatures reported from this diode are no more valid than temperatures reported from any other method of measurement. As long as all measurements from a set of data are taken from the same point using the same method, then they should still report equally valid results.

The noise level of the SSP under heavy load is tolerable but definitely not silent as the name of the cooler suggests. The 92mm fan, which must run at much higher RPM than 120mm fans to push the same airflow, is obviously the culprit here. However, the fan on the SSP runs far quieter than the 92mm fan on the Vigor Monsoon II, which ramps to 3500RPM (compared to 2500 on the SSP) and produces a sound akin to a jet engine.



Our overclocking experience with the SSP was an interesting one, and we went about it slightly differently than we have in the past. Our first overclocking attempt was at 3701MHz at settings we know to work flawlessly with the components we use. The system booted into Windows XP just fine and was very stable. After monitoring the temperatures at this speed for quite some time, we were convinced that this cooler would allow a bit more overclocking headroom. First, however, we wanted to see if we could get the processor to run at 3701MHz (285x13) at a lower voltage. The system would load Windows but fail whenever a benchmark was run to verify stability. We then bumped the voltage back up and went straight for 3750MHz, a departure from our normal practice of 15MHz (or 13 in this case) increments. When the system booted just fine with basically no increase in reported temperatures, we became pretty excited about the possibility of getting this chip into the 3.9-4.0GHz range. The next step was to 3850MHz, and the system booted just fine. After about 5 minutes of idle operation with relatively low temperatures, we received a blue screen. This was repeated several times until we lowered the clock speed to 3800MHz (293x13) for stable operation. We actually were able to play Team Fortress 2 for several hours at this clock speed to prove stability. The maximum overclock achieved on the Silent Square Pro is a full 100MHz beyond the previous best result. Pretty impressive to say the least.