State of single-GPU passthrough?

[the_jest]

I know this has been discussed extensively here, so apologies for asking again, but I can't tell if there have been any recent developments.

I've been planning on migrating from TrueNAS Core to SCALE, and was going to get a CPU with an iGPU so that I can use this for Plex transcoding. But I see now that it's only possible to pass through the GPU to apps if you have _two_ GPUs, so that one can be used by the system. Obviously, as many others have pointed out, I don't have a GPU now—it's a server—and I don't want to get another one that I don't need to use, just so that I can pass the original one along.

I see there have been various attempts to hack the middleware, add scripts, etc., to allow you to allocate your only GPU to an app. I'd rather not get into something this messy and fragile; I'd like this sensible and obvious need to be supported out of the box. What's the current state of affairs on this?

I'm feeling tempted to just get an older-generation CPU and mobo, and stay on Core, and then get a mini-PC to run a dedicated Plex server, if it's going to be this complicated....

 

[the_jest]

Bump, if that's OK (maybe I shouldn't have posted on Friday afternoon).

tl;dr: Just trying to find out if I really need to get 2 GPUs in order to be able to pass through 1 of them to an app.