Brumbaer / Computer Stuff / Macos on 9900k

Oktober 2018. Intel just released their new 8-core main stream processor, the 9900K.

Being an old Hackintosher it was just naturan to see whether it would run under macos.

The test system consists of

  • 9900K
  • ASRock Z390 Phantom Gaming itx/ac
  • 32 GB RAM in form of 2x 16GB G-Skill RAM. Frequency 3000MHz and latencies of 14-14-14-34.
  • Vega 64 Frontier.
  • BCM43602 compatible card in M.2 Adapter.
  • Samsung 860 EVO and Samsung 960 PRO SSDs.
  • EKB Liquid cooler, with two 280mm Radiators and a custom controller for pump and fans.

Putting the hardware together, was as straight forward as it gets, especially because I used this hardware before only with a 8700K instead of the 9900K.

To get macos running on a PC, you will need Clover. I install Clover and the Kexts needed manually and stay far awayfrom tools like Unitbeast and Multibeast. You never know what gets installed where, which is a pain if there are some problems. Also those tools change the macos System folder, which makes it difficult to use the system on different Hacks and makes trouble shooting harder. Manually installing Clover allows you to have all things needed to run macos on a PC in the EFI partition and leave your macos untouched.

As I used the motherboard already with the 8700K, it already ran under macos. It uses the same Clover configuration as it's predecessor board with a Z370. Even de XHC (read USB) controller is natively supported by macos Mojave. The only problem lay with the NVRAM, but were quickly overcome.

The cpu-family the 9900K belongs to ist also supported by Mojave, the only problem expected was the iGPU and so it was, it is not supported. For many functions you will need the iGPU running at least as a connector-less graphicscard.

That is easily be done by having the iGPU pose as a GPU like the one used in the 8700K. If you do not know already how to do this use Google.

I ran some benchmarks


In the meantime I did some more tests which changed the results slightly. I also gathered some comparison values for IMacPros. A video of the tests can be found here


Test9900K stockOC 5200iMacPro 10core14 core18 core
Luxmark Ball 4068 4192      
Cinebench OpenGL 171,03 190,32  139,50 144,99 145,19
Cinebench CPU 2061 2277 2026 2517 2950
Geekbench Single 6846 7022 5438 5135 5137
Geekbench Multi 38684 41645 36213 40478 46849


Luxmark Ball scales badly in comparison to the clock ratio. I checked for any throttling, but there seems to be none. It just doesn't use all of the cpus power. probably a RAM inflicted limitation.

The OpenGL scores scale perfectly which makes one wonder, because both tests use the same graphics card, which should be heavily used. Seems like the cpu is the bottleneck in this test.

As you can see the scores even of the stock 9900K is higher than that of an 18 core iMac Pro.

The CPU  Test scales well as it should be and the stock 9900K has about the same performance as the 10core iMacPro.

Geekbench Single scales well, even it doesn't look like it. The stock 9900K uses a higher cpu-clock if only one core is active, so the speed gain is smaller. The Multi test scales as expected.

Also as expected is that the Single scores a better than those of all iMacPros. Unexpected is that the overclocked 9900K scores better than the 14core.


The current overclock is set to 5.2 GHz and the RAM to 3.1GHz.

 The voltage is 1.39V fixed.

Temperatures depend on the operations performed.

Prime with a FFT length of 1344, draws 167W and reaches 62°C after 30 minutes.

Prime with a FFT length of 100, draws 217W and reaches 72°C after 30 minutes.

Prime with a FFT length of 80, draws 234W and reaches 78°C after 30 minutes.