Hacker News new | ask | show | jobs
by iso-logi 102 days ago
8 Core/16 Thread, boosting up to 5.1GHz with iGPU would be pretty neat for a Plex Server or Proxmox Server with a few VMs.
3 comments

As far as I can find, Plex does not support AMD iGPU for transcoding. Jellyfin will work, but support seems rather spotty. For other AI/ML work, it seems like ROCm is up and coming, but support - e.g. for Frigate object detection - is still a work in progress, especially for newer chips.
I have an AM4 AMD iGPU I use with Jellyfin; it works fine.
Is it actually using the iGPU, or just "brute forcing" it?

I've put it in quotes as the effort required from these chips for streaming transcoding is so low these days that brute force makes it sound like more effort than it really is.

>I've put it in quotes as the effort required from these chips for streaming transcoding is so low these days

What's your source for this? Transcoding without acceleration is incredibly expensive, especially for 4K content, and especially for 4K HDR content.

Even a single 4K HDR -> 1080p transcode takes a huge amount of resources.

The Asustor Lockerstor4 Gen3 has a Quad-Core Ryzen Embedded V3C14 and cannot transcode 4K content.

Meanwhile, an old Kaby Lake Intel chip does so just fine but only because its QSV can handle h265.

Thats interesting. My 5 year old Ryzen laptop can transcode 4k faster than realtime, which is what I mean mean by "these chips". Modern Ryzen, which is what the subject is about.

Quick Sync is invaluable for low powered processors, my old Intel embedded Wyze can do several streams.

"faster than realtime" doesn't mean much if it's in a device that's supposed to do more than just transcoding (such as serving a web app) or if you need multiple transcodes, etc.

Even on modern chips, transcoding is quite expensive.

People who are running Plex generally are running on servers also serving files, web apps, and who knows what else. These devices are often running 24/7, so both overall cost and power efficiency are big concerns. I wouldn't want to rely on my server being at high CPU usage most of the time - for power, heat, and overall reliability concerns.

Yes, it's actually using the iGPU. The CPU load remains very low.
Jellyfin supports it, but the resulting quality is noticeably poor compared to Intel QuickSync or software transcoding. Perhaps the newer chips are better, but if you're building a media server from scratch you'd probably build around an Intel CPU or ARC GPU anyway.
My proxmox server with a few VMs works perfectly fine with much less compute.

My homelab setup runs out of memory much faster than it does CPU cores.

Depends what you’re doing on the VMs, I run one as a desktop PC so have 4/6 cores and all the GPU access is important.
Maybe also Immich, for face and object recognition.