Test Results
Cooling Performance
The first thing we noticed is that the "spread" is pretty even here. Sure, there is a 16°C difference between the coolest and hottest load temperatures, but keep in mind that some coolers have two or three entries in the graph because of varying fan speeds. The absolute best result we see here under load is 73° C, and every other cooler in this comparison boasted at least one result within 11°C of that.
It's also interesting to see how differently the stock cooler performs depending on its fan speed. The stock cooler can perform the worst, or almost the best, depending on the fan speed you choose. Of course, while the stock cooler performs well at full fan speed, this comes at a terrible price as we'll see in the noise benchmarks later.
In fact, all of the solutions with variable fan speeds really offer large ranges of cooling performance, so with these coolers the tradeoff between noise and temperature is left at the user's discretion. In contrast, the one-speed-fits-all coolers like the Evercool Turbo 2 and the Thermalright V1 Ultra perform in the upper middle of the pack.
We can make a few overall observations from the thermal results. The KuFormula VF1 Plus manages to get the top spot, with the Thermalright V1 Ultra, Evercool Turbo 2, Accelero X2, and VF900-Cu close on its heels. The VF900-Cu seemed to perform slightly better than the competitor's coolers when the video card was at idle.
On a side note, there are those who would be concerned with the fact that the stock cooler dumps VGA-card-heated air out of the back of the case, but all of the aftermarket coolers in this comparison feed it back into the case. In our tests we were only able to detect a 4°C temperature difference in the case when comparing the stock cooler and the KuFormula VF1 Plus. While this isn't a huge difference, others have reported much larger temperature differences, because this is dependent on the air flow of each specific case. This is something the overclocking crowd should definitely keep in mind.
