UPDATED: CPU Cooler Testing (Thermals Explained)
There’s a fact, that not many of you may know and that’s that thermal testing of CPUs (well, anything for that matter) is a nightmare. I wont go into to details here as it may only serve to confuse, but what I will do is show you what the pcG Team did in an attempt to get our data as accurate as possible.
When we built our three Test Rigs for 2015, we first ensured that we all had the same components, the same fans speed and the same voltages etc. Once this was done we set about calibrating the three Test Rigs, and what we found was not surprising but it was interesting…
After performing our normal CPU thermal test using our Test Rig at 4.0GHz, that sees us running Prime95 (Small FFT) for 15 minutes and taking the average of the maximum Core temperatures, we found the following. Each Result was different, there was a total of 3.25 degrees between all of the Rigs. To get closer to the truth we decided (during another test) to try each CPU in the same Rig, rather surprisingly the difference was still there proving that not only is there a Silicon Lottery when buying a CPU, there’s a Thermal Lottery too!
With this in mind we have calibrated our Test Rigs; this means that when testing CPU Coolers the results from both Mike and Terry need to be adjusted (Calibrated). pcG Terry has a negative 1.5 Degree Delta and pcG Mike has a positive 1.5 degree Delta. This then brings both Test Rigs in-line with James’ Test Rig which is (effectively) the average.
UPDATE (12/01/15): Due to a new CPU being introduced these Deltas are now as follows: pcG Terry has a negative 0.75 Degree Delta, pcG Mike has a positive 1.0 degree Delta and pcG James now has a negative 0.5 degree Delta.
This hopefully explains the length that we have gone to in an attempt to give you the most accurate data we can, given our setup. If you want to be super accurate, you will need to apply these Deltas to our results, as we have not as it would only lead to confusion. Also remember that if you were to buy any or all of our Test components, you will also get a slightly different result, maybe even a better one! 😉
In an attempt to ensure that all equipment is validated we have invested in some new thermometers for ambient temperature readings. As you can see these three thermometers are all within 0.1 degrees of one another. This is plenty accurate enough for recording ambient temperatures.
FYI: The thermometers can be found on Amazon if you’re interested.