Cooling Efficiency Tests

Using the “weakest” cooling system with the fan in quiet mode we managed to overclock our quad-core processor to 3.88GHz without losing the system stability. The processor Vcore was increased to 1.5625V in the mainboard BIOS (to 1.53~1.54V according to the monitoring tools):

The results obtained in an open testbed are given on the diagram below:

Thermalright HR-01 X didn’t impress us with its performance and lost not only to SI-128 SE but also to IFX-14. The latter, however, demonstrated much more remarkable results. It proved as efficient as SI-128 SE with only one 120-mm fan running in quiet mode between the heatsink sections. However, once we installed two fans (one on the outside of the first section and another between the sections), IFX-14 started cooling 6ºC better! This is a really big difference considering that the noise level remained practically the same. The third fan taking the air away from the second section wins us another 2ºC of the CU temperature. And if we install one 120x120x38mm fan between the heatsink sections, the cooler will be 3ºC more efficient than with a  single 120x120x25mm fan. They have different blades rotation speeds, because we leveled them out to ensure they are equally noisy (~860RPMof Scythe SlipStream 120 are equal to ~1020RPM of Scythe Ultra Kaze).

Speaking of the noisy mode at maximum fan rotation speeds, IFX-14 with a single fan at ~2000RPM cools 2ºC better than SI-128 SE. And with a couple of powerful fans it gets 7ºC ahead of SI-128 SE. This performance difference as well as the coolers potential inspired us for a maximum CPU overclocking experiment, when we tried overclocking the test CPU to its maximum with each of the cooling solutions in each operational mode. Here are the results:

I would like to say that the difference during maximum overclocking of our quad-core processors turned out really small. It must be not just the overclocking potential of our CPU, but also high room temperature in the lab. Nevertheless, the difference between the two leaders, Thermalright SI-128 SE and IFX-14, is evident: sometimes it reaches 7ºC in favor of the super-cooler.

And what about the performance inside a closed system case? Here everything seems to be pretty simple, because Thermalright SI-128 SE offers absolutely unbeatable price-to-efficiency ratio. It yielded only 2ºC to IFX-14 with two fans at ~2000RPM (though we had to put up with very high noise level and remove three memory DIMMs). I don’t need a diagram is necessary here.

Of course, we haven’t forgotten about HR-10 heatsink. We tested IFX-14 with and without it. As a result, in the latter case the temperature of our overclocked CPU under maximum workload increased only by 2ºC. So, this result gives us every right to state that HR-10 does have certain effect on the CPU cooling. It is really a pity that this whole thing will not fit into any system case. As you probably understand, HR-10 installed on a different platform may demonstrate different improvement or turn out completely useless at all.

Thermalright HR-01 X

We also couldn’t leave out the efficiency of our cooling solutions in passive mode. That is why here are the results obtained after a 30-minute run in an open testbed (the room temperature during this part of our test session dropped to 23.5ºC, so it would be incorrect to compare these results with what has just been discussed):

Thermalright IFX-14

Well, Thermalright HR-01 X is not very impressive, I should say. Of course, it copes easily with a non-overclocked quad-core CPU. However, the processor temperature approaches 95ºC even during relatively modest overclocking to 3.46GHz and slight Vcore increase. As for its senior fellow, it totally amazed us with its efficiency in passive mode:

With a CPU in its nominal mode, Thermalright IFX-14 maintains its temperature at 58ºC, which is better than what a buzzing boxed cooler from Intel can offer. As for CPU overclocking with a passive Thermalright IFX-14, it coped just fine with a CPU at 3.75GHz with 1.4875V Vcore: the CPU temperatures never exceeded 86ºC even under maximum workload! We could have pushed the CPU frequency even higher than that (the throttling threshold was still pretty far ahead), but even this result is already very impressive. Neither Scythe Ninja Copper, nor gigantic Scythe Orochi could boast the same efficiency in passive mode.