CPU coolers

If your PC sounds like a hurricane combined with a tsunami then it doesn't matter how fast it is - it will be so uncomfortable to sit in front of that you simply won't want to use it for more than a few minutes.

The single noisiest component in your PC is likely to be the CPU heatsink and fan (HSF), so it makes sense to investigate the myriad of third-party coolers, many of which are significantly quieter than the reference Intel or AMD HSF supplied with your CPU.

However, cutting down on unwelcome fan noise isn't the only advantage to upgrading your CPU HSF, because a better HSF will also allow you to overclock your CPU further. As high-end CPUs, such as the Core 2 Extreme and Athlon 64 FX-series, are so damn expensive, a cheaper (and often better) option is to buy an entry-level CPU and overclock it. After all, a high-end CPU costs close to £600, but a new HSF could set you back as little as £12.

For these reasons, we've rounded up 22 different CPU HSFs, and ranked them in order of cooling performance, noise, ease of installation and value for money. Unless you're a masochist who enjoys noisy, hot-running PCs then the next few pages will be essential reading.

The single most important characteristic of a HSF is how effectively it cools your CPU in order to prevent it from throttling (reducing its frequency, and therefore its performance) or, worse still, automatically shutting down your PC to prevent damage.

For this Labs test, we've modified our testing procedures slightly to change how we measure the CPU temperature. Instead of using Asus Probe to measure the CPU temperature, which varies in accuracy from motherboard to motherboard, we used Core Temp. Core Temp is a freeware applet (www.thecoolest.zerobrains.com) that can read the temperature of the DTS (Digital Thermal Sensor) embedded in the core of every modern CPU. The DTS provides a much more accurate reading than any third-party temperature sensor, since its readings aren't affected by sloppy BIOS programming or a poorly calibrated external health-monitoring chip on the motherboard.

Even though Core Temp is capable of displaying the temperature of each CPU core, we think showing the relative temperature difference between each HSF is more important than showing the actual temperature. After all, what's really important is how much better or worse a third-party HSF is than the reference Intel and AMD HSFs. This is why we use the temperatures achieved when using these reference HSFs as the base level for our tests.

For this Labs test, we built two new thermal test rigs, which we'll use to test all cooling hardware we review in Custom PC. The Socket AM2 test rig comprises a 2.6GHz Athlon 64 X2 5000+ overclocked to 2.8GHz, with the HTT raised from 200MHz to 215MHz and the vcore boosted to 1.45V. The rest of the system comprises an Asus M2R32-MVP motherboard and a GeForce 7900 GS installed in a Cooler Master Stacker 830 chassis. The LGA775 test rig comprises a 2.13GHz Core 2 Quad Q6400 overclocked to 2.66GHz by increasing its FSB from 266MHz (1,066MHz effective) to 333MHz (1,333MHz effective), and the vcore to 1.4125V. This system also resides in a Cooler Master Stacker 830 chassis, along with an Asus P5B-E Plus motherboard and GeForce 7900 GS.

The temperature results that you see in the graphs below were taken once the CPU had reached its maximum stable temperature while running two instance of Folding@home on the Athlon 64 X2 test rig and four instances on the Core 2 Quad test rig. As folding is so CPU-intensive, this should be the highest temperature the CPU reaches. The temperatures displayed are relative to the temperatures achieved by the reference AMD and Intel HSFs.

Scoring

With each review, you'll see four separate scores, which are Cooling, Design, Value and Overall. The Cooling score shows how well each HSF cooled the processor compared with the reference designs, which scored an arbitrary 50 per cent. The Design score was calculated by awarding a certain number of points for each key design characteristic. For example, we penalised excessively noisy HSFs, and designs that exceed the AMD and Intel maximum weight recommendations.

We awarded points to HSFs that were easy to mount on the motherboard, and penalised those that were difficult to fit. HSFs that provide extra cooling for components around the CPU socket were also rewarded. The Value score was then derived by a weighted calculation of the Cooling and Design scores, divided by the price. The Overall score is also weighted, and not an equal average of the three initial scores. For this ratio, we used a weighting of Cooling (40 per cent), Design (30 per cent) and Value (30 per cent), since cooling is the most important characteristic of an HSF.