Cracking Hardware Myths

Тестова система

CPU: AMD FX-8120 @ 4.4 GHz
Memory: 2×4 GB Adata XPG G-series V2.0 @ DDR3-1866 10-11-10-28-1T
Motherboard: Asus Croshair V Formula, AMD 990FX
GPU: Radeon HD 7970
Hard Drives: Western Digital Caviar Blue 1000 GB, Western Digital Black 500GB
PSU: Seasonic X-760
Case: Cooler Master HAF 912+
Fans: 2x CoolerMaster Megaflow 200mm @ 750 RPM, AeroCool Shark 14 @ 850 RPM, Shark 12 @ 850 RPM.
Coolers: Noctua NH-U12P SE2, Corsair A70.
Display: Hannspree HF237, 1920х1080
OS: Windows 7 Home Premium 64-bit
Drivers: ATI Catalyst 11.12

The processor was overclocked to 4.4 GHz with voltage of 1.4625, which drops under load to 1.428 V. The frequency of the memory controller was left to its default settings. Application Power Management was disabled, and High Performance Mode – activated. I’ve tested NH-U12P SE2 with the standard fans at maximum speed and reduced to 900 RPM and also with reference Coolink SWiF2-120P at 1900 RPM and 1400 RPM. For comparison I used A70 equipped as standard fans operating at 2000 RPM and 1400 RPM and with Coolink SWiF2-120P in the above mentioned speeds. Fans in the case were left at the speeds mentioned above (maximum speed for MegaFlow, Shark models attached in lowering resistors), so that although the system is not completely silent, though the noise level is low enough.

The given results are the difference between room temperature and CPU temperature, the first during the tests was between 25.8 and 26,6 ^oC and was detected with digital room thermometer to the nearest tenth of a degree. The second was detected with AIDA64 with an accuracy of one degree. Furthermore, I decided to give the absolute values ​​detected, because my impressions are that such small differences in ambient temperature have no practical effect on the detected values ​​in the case. I used for both coolers Noctua NT-H1 thermal paste. Loading was done with Prime 95 26.5 x64 Blend mode for all cores, the temperature was detected in the 20th minute of the test. In idle mode the result was detected after 10 minutes. Tests were conducted twice and results – averaged.

Results

In idle there is nothing especially impressive as the A70 is in the lead, thanks to the direct contact and the faster spinning fans. NH-U12P SE2 falls slightly behind with standard fans at maximum speed and even further, when they work 900/min.

When processor is loaded, things do not change dramatically. In the end, however excellent workmanship, I have not believed that the Noctua NH-U12P SE2 has a chance against the Corsair A70. It has 40% more dissipation area, is using an 8 mm heat pipes instead of 6 mm and has much faster fans. And A70 beats at its maximum speed without any problems, leading by almost 5 degrees of absolute value. But things change when the fans rotates at approximately the same speed. In this case, somewhat unexpectedly for me Noctua’s model calmly confronts rival and even finishes slightly ahead of him. Reducing the speed of the fans of the NH-U12P SE2 leads to serious overheating of the CPU, although I was able to detect temperature, shortly thereafter always received an error in one of the cores. Which is normal for jump over 82 degrees Celsius. So these results are marked with a red background.

Equipping the two radiator with reference Coolink SWiF2-120P fans do not change things. Corsair A70 actually did not win more performance, while the Noctua NH-U12P SE2 wins 1 degree at a maximum and lost 1 at reduced speed compared to the two NF-P12 at 1300 RPM. It seems despite the small dissipation area, the main problem is not so dissipating the heat but the performance of the 4 heat pipes. Still impressive is the fact that such a radiator can manage with approximately 250 W of heat and deal with the consumption of a monster such as the Bulldozer chip.

Conslusion

Well, is the design of U12 infinitely old? I would say not by far. Though no longer on the very top, in the meantime it has won enough awards. Yes, against modern coolers it looses with few degrees, but this in no small part due to the difference in the fans. At the same time at similar fan speeds the performance is essentially identical to that of Corsair A70 – an excellent model of modern design. Along with that, noise is more than good – in fact you should have very quiet system fans to listen to it at maximum fan speed. So most of you can easily use the NH-U12P SE2 without dealing with speed control. The only drawback is its relatively high price – in Bulgaria it can be found for 112 leva incl. VAT, which is substantially above the price of Crosair A70 – about 80 levs incl. VAT. Along with that I wish that Noctua included in the kit the parts of the additional NM-A90 mounting system, but that is mostly of convenience for AMD users, rather than something very needed.

In the end, thanks to its still good performance, quiet fans and good compatibility with the relatively high memory (at least for Intel systems) it is worth noting Noctua NH-U12P SE2, especially if you are looking for primarily silent cooler. Thereof we give it our “Recommended” award.

Finally, it seems the concept of 2005 has still fruits to be harvested, since recently, the company has released an improved version of NH-U12DO A3, which already includes 6 heat pipes and 9 more fins. We can hope that it will soon appear on the market for desktop systems.

Pages: 1 2