-
Notifications
You must be signed in to change notification settings - Fork 24
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add nvidia libraries for hardware-accelerated en-/decoding #34
Comments
not sure how one would add this properly to the container image,
|
I could create a branch that builds onto the base image and adds in those libraries. I'm reluctant to add them to the master/pass branch images though purely due to size. Can I ask what is the rough size of those library files in total?
|
if i can help any further i would be happy to. |
It's likely that most of them aren't required. I see two listed in the logs you provided. They might have a few dependencies which you can see with |
libnvoptix.so.460.67 can surely be removed from the list, trims down overall space by 220mb already |
so it "seems" that we just need those 2 libs, just checked and plex sucessfully transcodes using HW acceleration with only those 2 files sizewize:
|
from ubuntu packages:
and the cuvid lib is in the corresponding decode package:
|
Great investigation! Thanks. I'll put something together later |
thanks so much, great work on all these spiritsail containers btw 👍 |
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
I've pushed some images to Docker Hub if you want to give them a try. The source for that image is on the Let me know if it works or not. I don't have Docker set up with nvidia yet. Last I saw it was a bit of a pain to try to get working. I might try setting it up if this image doesn't work for you. As for chat, I don't think we do really. You can always open a GitHub issue for problems (I'd always prefer these for tracking). You can email us @spritsail.io, the email addresses should be in the maintainer labels for each of the images. I guess you could always hit me up on Matrix |
Bump @psych0d0g |
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
ill check the build soon, sorry had a bit of private troubles lately that held me back from experimenting |
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
I stumbled across this tagged image looking for something similar. Would you be able to bump the pass version with these libs @frebib and I'll be another tester? |
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
@lwndow I'm 99.9% sure that this is broken since Plex moved to musl as it's libc, but I've rebased the I'm not sure how to make this work with musl as I expect nvidia don't provide anything other than the glibc-based libraries that Debian has. I briefly looked for source/upstream for these a while back but couldn't find it. Presumably this does still work with other images though so it'll require a little more work on my part Edit: I made it build, but it won't work, at least with the Debian-packaged libraries
|
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. This change is only a test and doesn't build the base Plex image first, which the CI pipeline should do when this is merged into pass. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
Pull the nvidia encoding and decoding libraries from Debian and include them in the image. This change is only a test and doesn't build the base Plex image first, which the CI pipeline should do when this is merged into pass. Fixes #34 Signed-off-by: Joe Groocock <[email protected]>
NVIDIA/nvidia-docker#509 (comment) Time to buy an AMD card I think 😁 |
DEFINITELY |
Is it at all possible to use a nvidia card with this container or do I need to go for the official pms-docker or linuxserver image? |
There's not much we can do here until nvidia either provide the source for their libraries (never going to happen) or they provide binaries compiled against musl instead of glibc. We could also bundle glibc into our image with these libraries which I think should work, but that's a bit of work to do. I'll accept a PR if anyone wants to give it a go. |
Please add librarys required for HW acceleration to work to the container.
The text was updated successfully, but these errors were encountered: