We place our test rigs under load using the Small FFTs stress test in the free application Prime95, as this test is designed to place maximum load on the CPU and minimum load on the system's memory. We use the freeware application Core Temp to measure the temperature of our CPUs, as it reads data from the DTS (digital thermal sensor) embedded inside each core of Intel's modern processors and thus provides as accurate a reading as possible. We always take the temperature of the hottest core for consistency, and do so after 15 minutes of the Prime95 test, which is enough time for temperatures to plateau.
The results provided are the delta T results; the difference between the recorded CPU temperature and ambient temperature at the time of testing. Our labs are not temperature controlled, so this provides the most accurate method of comparison for coolers tested on different days.
Click to enlarge - Never take the tundra lightly. It can drive a man insane.
Any cooler that causes one of our CPUs to hit its Tj Max, the maximum temperature at which the CPU will operate before downclocking itself, is considered a fail. If a cooler has multiple available fan speeds, we retest at appropriate speeds to highlight the differences. Unless otherwise stated, closed loop liquid coolers are installed with their fans as exhausts. Finally, any cooler that is PWM (pulse width modulation) capable is tested both at full speed and using our LGA1155 motherboard's standard PWM setting, with both results posted if there's a difference.
Intel LGA1155 Test System
Intel Core i7-2600K overclocked to 4.4GHz (44 x 100MHz) using a vcore of 1.275V
Asus Maximus V Formula
2 x 2GB Crucial Ballistix 1,600MHz low-profile RAM
PNY Performance 60GB SSD
Corsair HX750W PSU
Intel LGA2011 Test System
Intel Core i7-3960X Extreme Edition overclocked to 4.2GHz (37 x 113.5MHz) using a vcore of 1.275V
Asus P9X79 Pro
4 x 4GB G.Skill Ripjaw 1,6000MHz RAM
Samsung SSD 830 128GB
Enermax Revolution 87+ 750W PSU
Both systems run Windows 7 64-bit, are built inside a Corsair Carbide 500R and use an AMD ATI Radeon HD 5770 1GB. Coolers are tested with both of the case's side panels fully attached and with its two 120mm front intake fans and rear 120mm exhaust fan operating at consistent speeds, providing interior airflow that is more realistic than a benchtable or our previous set-up. The cases have been cable tidied and had their front drive cages removed to optimise airflow.
**Please note that we've now added results for certain liquid and air coolers that we've not yet had a chance to review on the site. The data comes from the CPU cooler Labs Test in issue 116 of Custom PC. All coolers were tested in the same systems and with the same methodology as outlined above.**