Skip to content

Conversation

JohannesGaessler
Copy link
Collaborator

This PR makes it so that CUDART_VERSION is printed in ggml_cuda_init if applicable. This information can be useful for debugging issues and I think it's easier to print it than to make users fill out issue templates.

@github-actions github-actions bot added Nvidia GPU Issues specific to Nvidia GPUs ggml changes relating to the ggml tensor library for machine learning labels Sep 7, 2025
@Green-Sky
Copy link
Collaborator

Would this make sense to print it even before initialization?

@JohannesGaessler
Copy link
Collaborator Author

I have yet to see issues where the program crashes prior to initialization so I don't think there would be much benefit.

Copy link
Collaborator

@Green-Sky Green-Sky left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I was thinking of linking/loading issues.

@slaren
Copy link
Member

slaren commented Sep 7, 2025

I would prefer this in ggml_backend_cuda_get_features instead, IMO it would be better to remove all the noise that some backends print on startup.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ggml changes relating to the ggml tensor library for machine learning Nvidia GPU Issues specific to Nvidia GPUs

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants