Cooling Efficiency Tests
First I would like to say a few words about the Swiftech MCP35X pump. Due to the fact that it supported PWM technology and also due to extensive functionality of the Intel DX79SI mainboard we could set the pump rotor speed at any desired value thus managing the noise levels as well as performance. Before moving on to the main round of tests, we studied the dependence of the cooling efficiency on Swiftech MCP35X pump performance and learned that the pump performance barely affects the efficiency of completely assembly systems. Judging by the pump rotor rotation speed readings, the cooling efficiency difference between the pump working at 1900-2000 RPM (very quietly) and 4000-4100 RPM (maximum speed, but tolerable noise level) was only 1°C! And most remarkably, this was true for all fan speeds of both radiators. Had there been any other components in this liquid-cooling contour, such as graphics card, chipset or voltage regulator water block, the dependence could be much more significant. So, I think it would be fair to say that the cooling efficiency of both Swiftech systems is minimally dependent on the pump performance provided there is only one Apogee HD water block in the cooling contour. Therefore, you can either use the PWM mode or set the pump speed at its minimum manually to ensure that it works in the most acoustically comfortable mode.
Let’s start with a warm up for the two new Swiftech systems: overclocking our processor to 4.375 GHz at 1.385 V Vcore. As usual, all results are summed up in the following diagram and table:
I believe no one is surprised to see the difference between two Swiftech liquid-cooling systems and Phanteks super-cooler: I doubt anyone had expected anything less from these products. Even at the lowest rotation speed of its two fans, the junior Swiftech H2O-220 Edge HD is 3°C more efficient than Phanteks PH-TC14PE cooler with the fans rotating at 800 RPM. And the top Swiftech H2O-320 Edge HD system turns out 8°C better under maximum load.
What we are indeed surprised with is the difference between the two Swiftech systems, because 1/3 larger radiator of the H2O-320 Edge HD allows it to cool our overclocked processor with all three fans at 800 RPM as efficiently as H2O-220 Edge HD would cool it with its two fans rotating at the maximum speed of 2060 RPM. In other words, if you have identical water blocks, pumps, tubing, coolant and fans just the size of the H2O-320 Edge HD radiator alone delivers 5°C better efficiency in quiet mode, and 6°C at maximum fan speed. And all that is with a moderately overclocked processor, nothing extreme yet.
As we continue, the efficiency difference will most likely increase even more. Let’s see what we get with a CPU overclocked to 4.5 GHz at 1.405 V Vcore:
True, as the heat dissipation of our six-core processor got a little higher, the senior Swiftech system is 5-7°C more efficient than the junior one at the same fan speeds. Moreover, a few things changed compared to the previous diagram: H2O-220 Edge HD at 2x2060 RPM was 2°C better than H2O-320 Edge HD at 3x800 RPM, and Phanteks PH-TC14PE at maximum and medium rotation speed of its fans turned out 1 and 2°C more efficient than H2O-220 Edge HD at 2x800 RPM. In other words, if we don’t take into consideration the noise and consider cooling only the processor, we can conclude that it won’t make much sense to replace Phanteks PH-TC14PE with Swiftech H2O-220 Edge HD. However, things are completely different from Swiftech H2O-320 Edge HD, as we can see.
Both liquid-cooling systems coped perfectly fine with an even more extremely overclocked processor: 4.625 GHz at 1.45 V Vcore. The results are summed up in the following diagram and table:
Phanteks PH-TC14PE managed to keep the CPU stable at this frequency only with both fans rotating at their maximum speed, while both Swiftech systems did perfectly fine in the entire supported speed range of their fans. What can I say here? The cooling efficiency of Swiftech H2O-X20 Edge HD systems is genuinely impressive.
For an even better illustration of the systems’ efficiency let’s add their results to our database diagram and table, where all coolers are entered in their default configurations. We will add the results in quiet mode and at maximum speed of their fans when the CPU was overclocked to 4.375 GHz at 1.385 V core voltage:
* - The peak temperature of the hottest CPU core is posted on the diagram taking
into account the difference from the current ambient temperature and is reduced to 25°C.
Even in the maximum CPU overclocking rating both Swiftech systems look great (especially in quiet modes):
Now let’s check out the level of noise generated by both Swiftech systems.