Thermalright Ultra-120 eXtreme 1366 RT CPU Cooler Review

CPU Thermal Testing @ Stock Settings

We are all ready to look at the results so let's just jump into it. First we will take a look at a standard "average temperature" chart of the results at stock settings, then we can further dissect the results with a line graph showing the temperature throughout the testing period.

At the stock settings of our i7 965 Extreme Edition processor, it is evident that the different fan configurations don't really influence the results to terribly much. The entire delta amongst fan setups of the averaged core temperatures is less than 3°C. Even with this small delta, a pattern of cooling hierarchy is starting to reveal itself. Push/pull seems to be the best performer with push coming a close second and pull being last by a larger percentage. Of course, all six fan configurations on the Thermalright Ultra-120 eXtreme clearly outpace the stock Intel cooler with the best TRUE configuration running 12% cooler.

Keep in mind, we also had the Intel CPU cooler running at the "performance" setting which means it ran even louder than the push/pull configuration of the high RPM ADDA fans. In reality it may not be any louder, but it certainly created a different pitch that we found more annoying than the low hum of the ADDA fans. As we have recently found out in the discussion thread of the EVGA X58 SLI review though, fan noise really is dependent on environment and the individuals judging.

As mentioned, we have decided to add a line graph plotting the complete temperature logging results throughout the testing. We can see the 60 second idle period after booting up, followed by the 20 minute load testing, and the final 5 minute idle period following load testing. This graph gives us a better idea how each setup initially handles the heat load from idle. It also outlines how quickly each heat sink recovers from the heat load at the end of the stress testing. There is nothing really out of line or interesting to discuss on the stock settings with the line graph here so let's move on to the testing at our overclocked settings.

CPU Thermal Testing @ OC Settings

The overclocked settings are going to generate a whole lot more heat for the heat sinks to dissipate. These i7 processors run rather warm, even at the stock settings, as we just saw. Increasing CPU voltage by almost 0.20v over stock by itself is going to create a lot more heat, but the PWM area will also be running a lot hotter which just means more heat in the CPU socket area for the heat sink and its fans to try and deal with.

What appears to be the most amazing thing about these results has nothing to do with the Thermalright Ultra-120 eXtreme, it is the fact that the stock 965 cooler stood up to the beating and actually ran the entire testing period. Of course our load testing only consists of 20 minutes, but we didn't think it would last 2 minutes let alone 20 without throttling. Turning our attention to the TRUE results, we can see the same pattern is present with the push/pull setups outperforming the push setups which in turn out perform the pull setups. Only this time, the higher RPM ADDA fans show a marked improvement over the stock Thermalright fan.

In fact, the ADDA push setup all but equals the average temperatures of the Thermalright push/pull setup with the Vantec stealth. Deciding which setup was quieter was debatable but both were about the same. This type of testing shows that we really do have a wide selection of fan choices with the Ultra-120 eXtreme, and the included Thermalright fan is a pretty solid performer at what we think is all but silent. The default push setup of the Thermalright fan lowers temperature by over 14°C or 15% from the stock Intel 965 cooler. In our testing environment, the included Thermalright fan is inaudible while the whine of the stock Intel cooler is rather annoying...not to mention we are not sure running an i7 at 90°C+ is a good idea for a 24/7 setup. Here is the same graph as above outlining the temperatures throughout the testing period.

The patterns shown here are about what we expected with a little more gap between the different results. The stock settings just didn't create enough heat to show a big difference between the setups, the overclocked settings do. We mentioned earlier that the CPU cooling was the only thing we were going to look at today. With the importance of PWM cooling on this motherboard, we also logged PWM temperatures during our testing, and we are going to look at those results next.