Cooler Master, DEEPCOOL, Noctua, and Thermaltake CPU Cooler RoundupWaco -
» Discuss this article (16)
Testing of the various CPU coolers will be accomplished installing the coolers into the test system mounted into a case, not a test bench. Most systems are built and mounted into a sealed (relatively) chassis, so this method will be used to generate the load and idle results to give a real world view as to what kind of cooling performance one can expect based on the test system listed below.
Of course, your results may vary, due to case design and ambient air temperature, by several degrees. The CPU load is generated by Prime 95 version 27.7 for a period of two hours with a cool down period of one hour after the computer has returned to an idle state. Real Temp 3.70 is used to log the temperatures over the time frame with the highest and lowest averages across the four cores of the Core i7 2600K test CPU. Ambient temperatures are kept at 22C throughout the testing to minimize the impact of a variable temperature. Each cooler is tested with the same thermal compound (Arctic Cooling MX-2) to keep conditions as similar as possible between coolers. While this makes the results vary a bit from the as-packaged CPU coolers, it will allow for a more fair comparison of the coolers in the same test bed and should give the best idea of their relative performance.
- Processors: Core i7 2600K @ stock & 4.4 GHz 100 x 44 1.35v
- Motherboard: Gigabyte Z68AP-D3 v1.0 BIOS F8
- Memory: 4x4 GB Mushkin Redline 997051 9-9-9-27 1866Mhz
- Video Card: XFX HD 7970 Black Edition
- Power Supply: Cooler Master Silent Pro Hybrid 850
- Hard Drive: Corsair Force GT 240GB
- Optical Drive: N/A
- Case: Corsair Obsidian 650D
- OS: Windows 7 Professional 64-bit
- Ambient Temperature: 22C
At stock speeds these coolers are all more than adequate to keep an i7 2600K cool and safe. In terms of noise, from most quiet to least quiet, the ranking is: Noctua NH-L12, DEEPCOOL Neptwin, DEEPCOOL Ice Blade Pro V2.0, Cooler Master TPC 812, and lastly the Thermaltake Water 2.0 Extreme. All of the coolers were run with their fans at full speed. The NH-L12 and Neptwin were very close to silent with the NH-L12 being just a bit closer to sound-less nirvana. The Ice Blade Pro V2.0 and TPC 812 both essentially tied in terms of noise, with both having a reasonably audible medium-pitched drone that was easy to pick out over the noise of the 650D case fans. The Water 2.0 Extreme, being mounted in the top of the case, was clearly audible when cranked up to full speed (as it was tested) but has the ability to be quiet when using the built-in fan controller.
With only 8 degrees C separating the pack under load there really isn't enough heat here to separate the boys from the men. Not surprisingly, the two water cooling units with 240mm radiators lead the pack with the TPC 812 and Neptwin in close pursuit. The Ice Blade Pro V2.0 doesn't fare poorly here, but it does hang near the back. The low profile design (along with the quiet fans) of the Noctua NH-L12 are enough to keep temperatures safe, but not enough to compete with the larger and louder coolers.
The real test is, of course, the overclocked settings. The i7 2600K at 4.4 GHz with 1.35 volts pumping through it is not a cool CPU and can easily overwhelm coolers that aren't up to the job. Even at idle, the coolers start to separate themselves by their relative performance, since all power-saving measures are disabled. Once I fired up Prime95, it became obvious which coolers could handle the heat and which could not. The new Water 2.0 Extreme matches the performance of the Corsair H100, albeit with less noise. Just behind the water coolers and ahead of the rest of the field, is the surprisingly good DEEPCOOL Neptwin. The TPC 812 and Ice Blade Pro V2.0 are on even ground here, but have jumped into the uncomfortable 80+C range that many wish to avoid like the plague. Again, trailing the rear of the pack is the NH-L12 low-profile cooler that just can't handle the heat that the 2600K can dish out at these speeds and voltages. It isn't quite in to dangerous territory, but at 87C in a relatively cool room, it is a lot higher than I'd like my CPU to be sitting for extended periods of time.