Noctua NH-U12P: Top Performance AND Silence

CPU Cooling Test Configuration

All tests use our new cooling test bed. This consists of a Rosewill R604-P-SL case sold by Newegg without a power supply. The Rosewill is typical of a moderately priced mid-tower case our readers might own. We chose this case because it is a Newegg top seller and includes a variable front intake louver and a quiet 120mm exhaust fan at the rear of the case. The case is also screw-less with components held in place by plastic holders instead of metal-to-metal connections. This appears to reduce case vibration and noise.

The power supply is a Corsair HX620W, which has proven in benchmarks to be an exceptionally quiet unit. The HX620W features a variable speed fan and a down-facing intake fan mounted just above the CPU space in the case. To eliminate the video card as a source of noise we have moved to a fanless card. Since we will move to Vista and DX10 in the very near future, the test bed runs an MSI NX8600 GTS that supports DX10 and cools with heatsinks and heatpipes. The reduced noise power supply and fanless video card help to dramatically lower system noise in the test bed.

The motherboard is an ASUS P5K Deluxe. This P35 chipset motherboard has exhibited outstanding overclocking capabilities in our testing. It can also mount the newest 1333 FSB Intel Core processors and can handle our existing high-speed DDR2 memory. The P5K3 uses heatsinks and heatpipes to cool board components so all motherboard cooling is passive. There are no active cooling fans to generate unwanted noise during testing.

The 120mm exhaust fan mounted to the rear of the case is below the system noise floor. We run that fan during performance and overclocking tests. However, system noise can be cumulative, so we turn off the exhaust fan during noise testing.

We run all cooling tests with the components mounted in our standard mid-tower case. The idle and stress temperature tests are run with the case closed and standing as it would in most home setups. Room temperature is measured before beginning the cooler tests and is maintained at 21C +/-1C (68F to 72F) for all testing.

For consistency of test results, we use a standard premium silver-colored thermal compound. In our experience, the thermal compound used makes little to no difference in cooling test results. This is particularly true now that processors ship with a large manufacturer-installed heatspreader. Our current test procedure uses this standard high-quality silver-colored thermal paste for all cooler reviews.

For comparison, we first tested the stock Intel air-cooler at standard X6800 speeds and measured the CPU temperature at idle. We then stress the CPU by running continuous loops of the Far Cry River demo. We repeat the same tests at the highest stable overclock we could achieve with the stock cooler. "Stable" in this case is the ability to handle our Far Cry looping for at least 30 minutes without crashing.

The same benchmarks are then run on the review cooler(s) at stock speed, 3.33GHz (10x333) at stock voltage, highest stock cooler OC speed (3.73GHz), and the highest OC that could be achieved in the same setup with the cooler being tested. This allows measurement of the cooling efficiency of the test unit compared to stock and the improvement in overclocking capabilities, if any, from using the test cooler.

We compare results with a representative sample of air- and water-cooling results measured with CoreTemp. TAT provides a similar core measurement, but test results with CoreTemp are more consistent over a wide range of test conditions than the results reported by TAT. We retested previously reviewed coolers with CoreTemp under idle and load conditions. In benchmarks where the new test bed makes no apparent difference, like maximum overclock, we include results for all coolers tested since beginning cooling reviews in early 2007.

Noise Levels

In addition to cooling efficiency and overclocking abilities, users shopping for CPU cooling solutions may also be interested in the noise levels of the cooling devices they are considering. We measure noise levels with the case on its side using a C.E.M. DT-8850 Sound Level meter. This meter allows accurate sound level measurements from 35b dB to 130 dB with a resolution of 0.1 dB and an accuracy of 1.5 dB. This is sufficient for our needs in these tests, as measurement starts at the level of a relatively quiet room. Our own test room, with all computers and fans turned off, has a room noise level that has been reduced slightly to 35.0 dB(A) compared to the previous 36.4 dB(A). With the new test bed, the system noise at idle is 36.5 dB(A) at 24" and 37.8 dB(A) at 6". This is better than our previous system noise floor of 38.3 dB(A) at 24". The noise reduction at the 6" distance is dramatically lower than the previous test bed floor of 47 dB(A).