Test System and Thermal Results
With the system listed above I apply Arctic Cooling MX-2 to all the coolers I have tested on the Gigabyte motherboard to even things out with all the cooler manufacturers. To gauge the idle temperatures I allow the PC to run for a few minutes from the fresh reboot. I then open RealTemp and let things again settle as the reading spikes until the processor is again fully idle. At this point I get the reading and report it to the charts. For the load testing, both stock and overclocked, we use Intel Burn Test to supply the system load. In this application I set it to work eight threads for 50 passes, and maximize the memory tested to over 3000MB to generate as much load as possible to the coolers. With RealTemp open for the entire run, I report the highest temperature seen during the test run. All cooler testing is done open air on a test bench with an ambient temperature of 24.5° - 25°C.
In addition with these tests, I incorporated the use of a fan controller and voltages were verified with a digital multi-meter at the fans 3-pin connection. Also prior to any testing, both of the new loops were run for a total of 2 hours each to be sure they were complete void of air in the loop to ensure the best results for the testing.
With all three kits, at idle, all three kits equaled out at the same levels. With only 0.9V running through the processor at idle, the results of twenty-five degrees with 12V to the fans and twenty-six degrees when supplying 7.5V, there isn't anything to complain about at this level as all three kits top the chart.
With the fans running at 12V, both of the Edge HD's beat the best air cooler by a minimum of five degrees, at stock levels. I moved the EPIC 180 up on purpose to show it right next to a custom loop versus the best AIO I have tested. The old Edge kit still beat the other air coolers, but fell to third behind both of the newer kits. That's right, even the new H20-220 Edge HD is more capable than its larger forerunner.
Once we overclocked the system, the gap between kits got a bit larger, but even here it was only a four degree spread across the three Swiftech entries. At 12V the 320 HD just edges out the 220HD by a degree, and both are three degrees cooler than the original. What I really like to see here is that the radiators are almost as efficient for just a CPU loop to handle everything at 7.5V to the fans. Once you add the additional loops to the HD system, the 12V may show more advantage, but for just a CPU, turn the fans down and enjoy the low noise, much cooler CPU cooling solution.