Adding CUDA release for Ubuntu #16773
Open
+74
−0
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hello @slaren,
I hope this finds you well.
This release script alteration adds a job that results in two artifacts:
llama-bin-ubuntu-cuda-12.4-x64.zip (the llama.cpp library compiled with cuda support)
cudart-llama-bin-ubuntu-cuda-12.4-x64.zip (stand-alone version that saves the trouble of having to download any third part dependencies saving about 7 GB of dependency downloads and the compilation time)
Tested on Ubuntu 22.04 x64.
I wrote the job with the same conventions as the windows CUDA versions and the Ubuntu release jobs in the release file.
The result is a package exactly like the cpu version with 1 extra file (libggml-cuda.so) in case of the first package or 4 in case of the standalone one (libggml-cuda.so, libcublas.so.12, libcublasLt.so.12, libcudart.so.12)
I only used the Ubuntu and Nvidia officia packages for the build no other third-parties. In other words, it will work for all Ubuntu 22.04 x64 systems.
Best regards.