oneVPL-intel-gpu icon indicating copy to clipboard operation
oneVPL-intel-gpu copied to clipboard

[Bug]:

Open maxvaneck opened this issue 7 months ago • 1 comments

Which component impacted?

Encode

Is it regression? Good in old configuration?

None

What happened?

while trying to run both ai and av1 transcoding on my arc a750 i filled up the memory (unrealistic workloud i know)apparently jellyfin-ffmpeg doesn't like that and and just simply segfaults .

Image

here's the ffmpeg logs

logs(2).txt

the folks over there decided this was an intel problem and pointed me here

this happens on i fill the memoryusing koblodcpp which uses llamacpp as a backend everything works just fine if i don't do this

Platform: unraid linux 6.12.24-Unraid ffmpeg version 6.0.1-Jellyfin

** unraid 7.1.4

Virtualization: Docker container: FileFlows

Hardware Acceleration: QSV

CPU & GPU Model: cpu = i5 11400 gpu = intel arc a750

i'm happy to provide additional information but i'm a bit limited as to what unraid can do

What's the usage scenario when you are seeing the problem?

Transcode for media delivery

What impacted?

No response

Debug Information

No response

Do you want to contribute a patch to fix the issue?

None

maxvaneck avatar Jun 22 '25 10:06 maxvaneck