These components are always present in a computer system, so while FrostyTech is able to provide comparative data by removing the CPU from the system, manufacturers must vet the placement of temperature sensing thermistors, and other apparatus before a heatsink comes to market. In some systems with poor ventilation, the ambient case air temperature can rise as high as 50C, and that can certainly make a serious impact on a CPU heatsink.

Four thermocouples are arranged over the top of the heatsink fan with Kapton tape, they record ambient air temperature as it relates to the heatsink. The placement can very from one test configuration to the next, but so long as consistent placement is maintained the final result will be representative.

With the system running, the case temperature will rise and a manufacturer like Thermal Integration can look over the data and determine if the product has what it takes to maintain safe CPU operating temperatures. This is how heatsink makers test heatsinks according to Intel Thermal Design guidelines.

Not all heatsink manufacturers do things the same way as there is more than one method to choose from when evaluating heatsink effectiveness. Elsewhere at Computex in Taipei, FrostyTech stopped by the Thermalright exhibit where that manufacturer had an altogether different method of testing its heatsinks on display.

The Thermalright platform was altogether more akin to those used by several hardware publications; a set of variable DC power supplies provided current to a resistive heater below some manner of metal 'CPU die' housed in a phenolic test platform. A pair of hand held thermocouple thermometers record the die and ambient temperature, and based on those values C/W can be calculated. In the example above, a Coolermaster heatsink was being put through its paces on the CPU test stage.