Testing and benchmarking
Now that the system is hooked up to the active and actually working cooling solution let's run some performance tests. Initial thoughts: very impressive. The CPU is running steady at a ~22-23 C temperature right after windows booted up. You'd almost think 'naah, that's not right'. You simply are not used to temperatures like these with the new systems. As part of the test we maxed the workload of the CPU to 100%, managing some serious encoding work.
The funny part is the temperature does not rise that much with a full 100% workload, it only differed a couple of degrees but basically that's it. This creates some serious overclocking potential for the CPU.
Cooling method versus temperature in degrees C. The lower the better.
When we do temperature tests we always set the room temperature at 21 Degrees C. As you can see the CPU in idle is at an all-time low 22-23 Degrees Celsius. With traditional AIR cooling using the stock cooler we were at roughly 40 Degrees C. Giving the CPU a little edge immediately results in a slight higher temperature. It never ever once exceeded 35 Degrees C though and for a passively cooled system that is a nice accomplishment. Once the CPU is not stressed anymore within seconds the temperature retunes to very low values. This means that the waterblock can absorb lots of heat and the reservoir can cool down the fluid quickly.
The thing that is most interesting is of course the temperature when we give the CPU a looped burn-in test. It'll work at its hardest thus at 100%, now get this it never exceeded a temperature of 35 Degrees C, the same as most high-performance watercooling kits.
When we look at the AIR temperature at 100% utilization, we noticed 52 Degrees C. That's quite a difference eh? When we relate this to overclocking, your results are going to be limited by the model of CPU or graphics card that you have. Therefore we won't as our overclock results would not tell you the slightest thing.
Here is a small overview of the more recent watercooling kits we tested. The score is of course temperature in degrees C, the lower the better.
When you look closely at these results you need to bare in mind that the 2.8 Ghz Pentium 4 and Athlon 2000+ processors of course were much more cooler processors then the newer CPUs. I find the results really wonderful for this kit.