I don't wanna ruin everyone's RadeonBoost party, but so far in my testing, I only found my system to be slower with the kext. Let me explain:
I'm running a VEGA 64 8GB and 9900KS on Clover with iMacPro1,1 with iGPU disabled (in hindsight, the 9900KS doesn't make any sense whatsoever). I tried various setups: iMac19,1 instead of iMacPro, enabled and disabling iGPU, with and without the SSDT. Even though I do sometimes see a significant increase in GB5 scores, I have never experienced an increase in render times in DaVinci Resovle 16. Futhermore, all tested setups and combinations resulted in significantly lower performance. I tested the same video with each setup (a combination of multiple camera's and formats including H264, H265, retiming, LUTs and titles). For instance, using the setup as described above, without RadeonBoost, gets me about 55 fps when rendering 4K H264 clip. Using RB, this drops to around 30-35. As video rendering and video editing is the main reason I built this setup, this kext unfortunately doesn't make sense to me.
I think it's important to talk about this. Better GB scores are nice and all, but translation to real-world performance is way more important to me.