immich_machine_learning+openvino is using my intel cpu's igpu instead of my arc a770 dgpu
Fresh install via docker on ubuntu 24.04. using intel_gpu_top shows my 12th gen i5 igpu is being utilized at 100% while my 16gb A770 card is sitting idle. Is there some way to force the device selection? I was not able to find anyone else having this issue...
33 Replies
:wave: Hey @Josh at WLTechBlog,
Thanks for reaching out to us. Please carefully read this message and follow the recommended actions. This will help us be more effective in our support effort and leave more time for building Immich :immich:.
References
- Container Logs:
docker compose logs docs
- Container Status: docker ps -a docs
- Reverse Proxy: https://immich.app/docs/administration/reverse-proxy
- Code Formatting https://support.discord.com/hc/en-us/articles/210298617-Markdown-Text-101-Chat-Formatting-Bold-Italic-Underline#h_01GY0DAKGXDEHE263BCAYEGFJA
Checklist
I have...
1. :blue_square: verified I'm on the latest release(note that mobile app releases may take some time).
2. :blue_square: read applicable release notes.
3. :blue_square: reviewed the FAQs for known issues.
4. :blue_square: reviewed Github for known issues.
5. :blue_square: tried accessing Immich via local ip (without a custom reverse proxy).
6. :blue_square: uploaded the relevant information (see below).
7. :blue_square: tried an incognito window, disabled extensions, cleared mobile app cache, logged out and back in, different browsers, etc. as applicable
(an item can be marked as "complete" by reacting with the appropriate number)
Information
In order to be able to effectively help you, we need you to provide clear information to show what the problem is. The exact details needed vary per case, but here is a list of things to consider:
- Your docker-compose.yml and .env files.
- Logs from all the containers and their status (see above).
- All the troubleshooting steps you've tried so far.
- Any recent changes you've made to Immich or your system.
- Details about your system (both software/OS and hardware).
- Details about your storage (filesystems, type of disks, output of commands like fdisk -l and df -h).
- The version of the Immich server, mobile app, and other relevant pieces.
- Any other information that you think might be relevant.
Please paste files and logs with proper code formatting, and especially avoid blurry screenshots.
Without the right information we can't work out what the problem is. Help us help you ;)
If this ticket can be closed you can use the /close command, and re-open it later if needed.compose file, env file etc please 🙂
This is all managed through those
lspci:
You have drivers for this card installed, yes?
yes I use it with locally hosted LLMs and generators
What does
docker exec -it immich_server /bin/bash
And then ls -la /dev/dri tell you?mmm two cards
does your server have an iGPU?
yes I guess the title didn't carry through to the discussion
immich_machine_learning+openvino is using my intel cpu's igpu instead of my arc a770 dgpuits just a regular i5
model name : 13th Gen Intel(R) Core(TM) i5-13400
ah it did, I just didn't see it 😛
no worries I appreciate your help
I think the quickest solution would be mounting only the Arc one
would it be via editing
to something like
Yes exactly, either in the hwaccel/service files or directly in the compose
I guess the extra files are easier to manage 😛
I figured I'd follow the install guide to reduce any possible issues 🙂
That's perfect yes, usually it's the opposite 😅
Right @Josh at WLTechBlog you'll also need to add the renderD129
It's obvious to me but I forgot not everyone is used to this
card1 + renderD128
card2 + renderD129
yeah I was just coming back with "no worky" 🙂
let me fix that
ok now I see what I want to see in intel_gpu_top 🙂
Excellent
don't forget the ML container
that's the ml container
Your machine learning jobs will go 🚀
oh 😄
i need to do the transcode still
I thought transcodes
log indicates transcodes are not being accelerated
immich_server | [Nest] 7 - 12/26/2024, 3:25:27 PM LOG [Microservices:MediaService] Transcoding video c7ff9cd5-d80d-49bc-9ece-575965374bc1 without hardware acceleration
And you've just changed:
to
I take it?
root 45998 429 0.2 1765768 375220 ? SNl 15:27 1:12 /usr/bin/ffmpeg -i upload/upload/4c20c761-7ac1-498e-9664-e5556d3d95e2/39/a9/39a9df20-a93f-4f27-992a-11138da8250a.mp4 -y -c:v h264 -c:a copy -movflags faststart -fps_mode passthrough -map 0:0 -map 0:1 -v verbose -vf scale=-2:720 -preset ultrafast -crf 23 upload/encoded-video/4c20c761-7ac1-498e-9664-e5556d3d95e2/ed/69/ed69c0af-45d4-419e-afd3-94270dc3a767.mp4
I'm not super duper well versed in this
but maybe try adding the render/video groups to the container
getent group render
getent group video
And then
replace 122 and 123 with whatever the numbers are
(Shouldn't make a difference but you never know)i'm browsig the source a bit to see if I can determine how the ffmpeg cli is being built
aha... it isn't enabled by default, you have to set it in the admin ui!
immich_server | [Nest] 7 - 12/26/2024, 3:35:34 PM LOG [Microservices:MediaService] Transcoding video 68c07904-c9f5-4500-bb76-991ef18dfe5a with QSV-accelerated encoding and decodingvery very nice now
ah yeah of course 🙃
If you have any further questions let us know, otherwise you can just /close the thread
thank you so much for your assistance
This thread has been closed. To re-open, use the button below.