Test System Setup
I would first like to thank ASUS, InWin, Corsair, and Fractal Design for supplying products for me to test with.
To test our video card coolers, we use the test system and software seen in the chart above. After the system has been booted and has idled out to its lowest temperature, we open up GPU-Z and after a few minutes we take a reading of the lowest temperature shown for the idle results. This is for both stock and aftermarket coolers. To provide the load for our testing, we have chosen Unigine Heaven as our heat source. We run the test at default settings, only changing the resolution to 1080p. After running in benchmark mode for twenty minutes, we again go to GPU-Z for results. To make sure we are getting the correct readings, not only do we log the results, but we also set the sensors to keep the maximum values shown, so when the software is closed, we can verify with the logs on the maximum temperatures for the various runs and thermal applications for each cooler.
The idle temperatures are the best we have seen, even against the full cover water blocks, but there is less to cool with the Kraken G10, as it only cools the GPU die with liquid, so it's not really a surprise either. At this point, it appears all the time and energy taken to get this kit assembled for testing might be worth it.
For points of reference, the 92mm fan was spinning at 1500 RPM, and the pump on the X40 was running full speed for all testing, but we did limit the fan to 7.5V of power, and had it spinning at only 1100 RPM at this point.
When the load is applied, the full coverage water block still takes the top of the charts, but there is much more cost involved with that test setup than the Kraken X40 plus the Kraken G10 kit, and being only two degrees behind the best on the chart is quite the accomplishment.
Once the overclock is added in, this kit really shines, and it only increases a single degree. We do have the pump still spinning at full speed, and the G10 fan is still cranking at 1500 RPM, but this time the X40 fan was allowed to spool up to full speed at 1900 RPM.
Two other things to cover are the noise, and reports of the cards heating up under the fan. First of all, the noise level of the Kraken G20 is a steady 33dB from the fan. However, the X40 runs at 37dB at idle and roared up to 52dB when powered with 12V. So if noise is an issue, choose wisely as to which AIO is going to work for your needs. Secondly, on the reports of cards getting hot, without heat sinks on the power delivery system, it is obviously going to heat. We got lucky on two points. Our card is very efficient, and even with an IR thermometer we saw two degrees difference between the stock cooler and our Kraken G10 kit. Just be sure you know what sort of heat you have to deal with, plan accordingly, and issues like that will be kept at bay.