setting aside a portion of system resources for jellyfin?

[JKAMAN280]

Hello everyone!
So I am looking to setup a jellyfin server and am contemplating having it run off of my personal desktop. Given that I have 2 2080tis on my desktop, is there a way that i can dedicate the second one souly for jellyfin, that way if I am running any sort of applications, they aren't affecting the performance of the jellyfin server, and transcoding isn't going to effect the performance of my desktop. Currently running windows 10, although id be open to running a debian based version of linux (probably mint). Ultimately the goal is to have it automatically run when I start up the pc, and keep everything separate in terms of software so they don't effect the performance of one another. Im also contemplating automating some processes in terms of organizing media, although ill worry about that once I actually have the server up and running.

Current system specs:
ryzen 7 3700x
32 gigs of ddr4 ram running at 3600MHz
2 founders edition 2080tis
1tb nvme for boot drive
2tb hdd for games
eventually 2 14tb hdds for the jellyfin server (possibly run them in a raid configuration to save me the headache if one of them fails)
all cooled by a custom water cooling setup so noise/temps arent to big of a issue

Anybody have any experience with such a setup?
Thank You!

[TheDreadPirate]

If you were to run Linux and then use Docker, Docker has a mechanism to pass in specific GPUs.

[JKAMAN280]

If you were to run Linux and then use Docker, Docker has a mechanism to pass in specific GPUs.

Is it possible to do this in windows? (only reason i ask is because nvidia drivers on linux aren't the greatest)
However if running linux is the best way to do it than ill just do that.

[TheDreadPirate]

Outside of Docker/LXCs/VMs, even just running Jellyfin directly on Linux, I am not aware of a way to assign Nvidia GPUs.