Cooling Performance
For cooling testing, I first just have to point out that our new test suite is throwing most air cooling options in with the wolves. Our new testbench is running the 12900K which thermally puts out a TON of heat. This is a worst case scenario and it is hard to keep cool with ANY cooling option. But it does push the two 120mm heatsinks to their limits. So to start things off we have the real worst case scenario test which uses AIDA64’s Stress Test using the FPU workload which puts a workload that is more demanding than everyday tasks. This is basically like rendering. All of the AIO coolers tested struggled with their huge dual 120mm and dual 140mm radiators and with the stock fan profile none of them kept the i9-12900K under 90c. The NH-U12A ended up landing at 97c and the NH-U12S pegged at 100c. While this was too much for both coolers, it does show that there is a big difference between the two 120mm Noctua designs. Turning the fan to 100% didn’t do anything with the NH-U12A because it was already running at that with the stock profile so the results are the same for both.
AIDA64 FPU Stress Test |
Stock Fan Profile |
100% Fan Speed |
Corsair iCUE H100i ELITE LCD Liquid CPU Cooler |
93c |
90c |
MSI MAG CoreLiquid C280 |
91c |
90c |
Sapphire Nitro+ S240-A |
91c |
89c |
Cooler Master MasterLiquid PL240 Flux |
90c |
89c |
Noctua NH-U12A chromax.black |
97c |
97c |
Noctua NH-U12S |
100c |
100c |
Using that same AIDA64 FPU Stress Test I also went back in and uncapped our PL2 settings. This time only focusing on the wattage being pulled shown in HWMonitor. The goal here is to see how much overclocking each cooler can offer but with the NH-U12S it also helped show the gap between it and the NH-U12A. The NH-U12A was stuck right at 250 watts which is also the same wattage that is suggested for the 12900K. The NH-U12S came in under that with it underclocking the CPU down to 233 watts. The AIO coolers did offer a little more here, but I was surprised that the NH-U12A was able to handle the 12900K with the FPU workload without it underclocking it under the official recommended wattage.
AIDA64 FPU Stress Test With PL2 uncapped and 100% Fan Speed |
CPU Wattage |
Corsair iCUE H100i ELITE LCD Liquid CPU Cooler |
262 |
MSI MAG CoreLiquid C280 |
272 |
Sapphire Nitro+ S240-A |
258 |
Cooler Master MasterLiquid PL240 Flux |
261 |
Noctua NH-U12A chromax.black |
250 |
Noctua NH-U12S |
233 |
Now the AIDA64 Stress Test with the “CPU” workload is a lot more like the normal load you will put your system on. This is comparable to gaming, browsing the internet, and other normal use. This also shows that when you aren’t running crazy workloads like rendering the air coolers can hold their own. The NH-U12A came in at 70c which was 2c higher than the H100i and 5c under the NH-U12S. Turning the fan all the way up was similar with the NH-U12A being within just 1c with two of the AIO’s and 5c above the best performing ones.
AIDA64 CPU Stress Test |
Stock Fan Profile |
100% Fan Speed |
Corsair iCUE H100i ELITE LCD Liquid CPU Cooler |
68c |
62c |
MSI MAG CoreLiquid C280 |
63c |
62c |
Sapphire Nitro+ S240-A |
65c |
57c |
Cooler Master MasterLiquid PL240 Flux |
63c |
57c |
Noctua NH-U12A chromax.black |
70c |
63c |
Noctua NH-U12S |
75c |
70c |
So like I mentioned before, the FPU workload is a lot like rendering. I did also use Blender to do some rendering and while I can’t run the render as long as I can with AIDA64 you can see the results are similar.
Blender Stress Test |
Stock Fan Profile |
100% Fan Speed |
Corsair iCUE H100i ELITE LCD Liquid CPU Cooler |
90c |
88c |
MSI MAG CoreLiquid C280 |
90c |
89c |
Sapphire Nitro+ S240-A |
86c |
86c |
Cooler Master MasterLiquid PL240 Flux |
83c |
89c |
Noctua NH-U12A chromax.black |
96c |
96c |
Noctua NH-U12S |
100c |
100c |