Swiftech H220 AIO Review
Intel i7 3930k
Stock@ 1.1v (undervolted)
4.0GHz @ 1.25v
4.4GHz @ 1.35v
4.6GHz @ 1.45v
Gigabyte X79 UD3
Corsair Vengeance LP Memory
Corsair HX850 V2
Corsair Force GT 60GB
Coolermaster Storm Trooper
So we have arranged our tests to gradually get more and more strenuous, with a 2011 system able to put out a huge amount of heat we needed a basic test for basic coolers. Most CPU's at stock don't need stock volts and many people don't realise this. We have managed to get our 3930K running at stock at just 1.1v. This is 100% stable and will be our initial baseline test.
We run all of our tests in the same way, only the fan speeds and the overclocks differ. We use OCCT in Linpack X64, AVX compatible with all logical cores tested and 90% free memory utilised. We have set the test up so its simply a click and run affair, there is a 10minute idle period for the system to settle followed by a 30minute torture test. We have allowed a 5 minute cool down to the end of the test to round the total test time to 45minutes.
We have set an upper limit of 80C as a maximum, OCCT has a feature that it will stop the test if temperatures exceed this limit. During testing we have noticed that it fails if one of the cores hits 82C. Never the less if the program halts the test we consider this a failed run. In all of the H220 tests both the pump and the fans were run @12v.
Moving on to the 4GHz test, this is at 1.25v which is pretty close to what boards run at when left on 'auto'. If you are a regular on our forums you'll know how many times we end up telling beginners auto DOES NOT mean stock when overclocking!!! This does make quite a large difference in temperatures though with over a 10c increasing being seen on most coolers over our stock test. This is a good comparison for a basic overclock.
Our 4.4GHz run has the volts increased again to 1.35v, this does put a lot of demand on the cooler and only good performing heatsinks should make it into this graph with decent temps. You can see the D14 and the H100 both dipping there toes into the 70C averages here showing that they are getting close to the limits of our acceptable maximum.
Finally our 4.6GHz test. Don't be fooled, this is an extreme test and the graph reflects this, you will only see the cream of the crop featured in this graph. The most surprising result here for us was that the H80i passed this test, quite the achievement considering both the 140mm units, the H100 and the NH-D14 all failed.