Performance
The performance of the NH-L9a-AM4 has been tested on the same platform as most other AM4 coolers, so the 3700X CPU and ASRock X570 Extreme4 motherboard.
As a side note, in our tests, we had similar results on the 3700X and the 3600/3600X even though there is a two-core difference, and the 3600X is 95W TDP processor while others are 65W TDP.
Our comparison includes three tests. The idle mode is a PC left without any load besides standard Windows services running in the background. The mixed-mode base on a PCMark 10 extended test that uses popular applications and simple games. The max load is a CPU+FPU AIDA64 stability test. Shows about maximum CPU load during the most demanding work on all CPU cores.
The same as on the NH-L9i, our maximum temperature is high, but in the case of the NH-L9a-AM4, it’s slightly below throttling point. The NH-L9i had about 1-2°C worse maximum temperature. I guess it’s because of the slightly smaller heatsink as it’s designed mostly for Intel processors. The NH-L9a-AM4 is designed for AMD AM4 socket, so it can be a bit larger and will still fit every motherboard.
The NH-L9a-AM4 is slightly better than the stock AMD cooler, but also maximum boost frequency is higher, and the CPU is not throttling for most of the time. I won’t say it doesn’t throttle at all, but it’s barely visible while on much larger AMD Prism cooler it’s pretty common to see 100-200MHz CPU clock drops.
The generated noise has been measured about 1m from the PC in which was installed the NH-L9a-AM4. These results are similar to what we could see on the NH-L9i. While idle, the fan was spinning at about 1000-1200 RPM and was generating noise up to around 26dB. This is already quiet. During mixed-load tests, we could see between 1200-2000 RPM, and depends on the test; we could clearly hear it. The results were around 32dB. The maximum load made the fan to work at 2300-2400 RPM with the result of 37dB. It’s not quiet but pretty good result considering that the cooler was tested on an eight-core CPU, which has real TDP much above specified.
Because of the used 3700X processor, overclocking tests are not really possible. Another thing is that Ryzen 3000 series processors are not really overclocking, so we can say that optimal results can be achieved without overclocking. There are, of course, exceptions, but for overclocking, I recommend higher Noctua coolers, which are great for every platform.
