While I understand Open Core Legacy Patcher is not supported, with a MacPro 5,1 and RX580 GPU, Photo AI works correctly but while reporting it detects the GPU clearly the GPU is not utilized where the work load is pushed to the Xeon. AI Processor set to auto or RX580.
That said Geekbench 6 clearly is able to use the GPU with both Metal or OpenCL.
I’m curious if there is fix to get the app to use the GPU for processing.
Topaz Photo AI [v1.3.12] on [Mac Ventura 13.4.1 with OCLP 0.6.7]
I think your question is more suited to one of the hackintosh forums not here.
A MacPro is not a hackintosh and to be blunt there are more than a few photographers still hanging onto their MacPro 5,1’s in the exact configuration I’m running…
I know exactly what opencore is as I was one the first to test it back in the day while clover was still used. And I can see what legacy patcher is doing.
Your still patching the kernel therefore “Hacking” and I doubt no one here would be able to help.
I have changed this to the General category as it isn’t a bug. This is an unsupported platform but there may still be users here that can help you.
While not definitive I believe I understand what is going on between running Photo AI (and other Topaz AI tools) on MacOS version 12 versus version 13.
In short Topaz leans on Apples CoreML where OpenML makes vague claims of support CPU (GPP), GPU and neural processing subsystems. Clearly in version 13 of the OS Openvino is utilized. Openvino in fact targets Intel CPU instructions and GPU hardware, not AMD.
My sense, though Apple is vague on hardware requirements, is that in version 12 support was still present of AMD GPU hardware. But in version 13 of the OS AMD GPU support was dropped in favor of a library which specifically targets Intel hardware.
Just now referring to Topaz’s system requirements only very top level statements are made for Intel CPU or Apple M1/M2, where no reference is made to GPU. Likely it would be useful to other users for Topaz to further refine their system requirements.
With respect to the troll in this thread, this has little to nothing to do with OCLP. My gut says that for the remaining Intel based Apple platforms that are still supported in version 13 of MacOS (Ventura), those systems if they have AMD GPUs will also come up short and be forced to use the CPU (GPP) for ML processing.
Let’s just have some perspective here shall we.
You have only just installed Ventura 13.4.1 in the last 24 hours on your 2010 MacPro, you also failed to state Topaz was not the only software you are having with this problem.
I was just pointing out that using a boot loader to inject and patch data in memory to run an unsupported operating system (Mojave 2018 was last official supported OS for mac 5.1) is way beyond anyone’s understanding on these forums.
Judging by the amount of people using OCLP and having serious issues running Ventura, I’d say it has a lot to do with running a stable / usable system.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.
