Create/update new dcgm tracks

This can be a little confusing so I’ll try to be as clear as possible.

Initially we asked to create tracks based in the base that the snap was created because we thought that dcgm snap could stop working because of mismatch of kernel and drivers versions. See this request.

In the end the issue that we initially found was because other reasons. The base that the snap is built is not affecting functionality and even on older Ubuntu releases like focal and jammy, the snap is working as expected

With the release of dcgm v4, we need to change the tracks.

What is now core24 track we want to change to v3 if possible. If changing name is not possible we want to create a new track v3

core24 previously asked to be created can be deleted or changed to one of this new tracks:

  • v4-cuda11
  • v4-cuda12
  • v4-cuda13

Summary

Creation of the following channels:

  • v3
  • v4-cuda11
  • v4-cuda12
  • v4-cuda13

If renaming core24 to v3 is not possible or can cause issues, creating a new track v3 is fine

Rational

On v3 dcgm used to ship all the necessary cuda compatible libs for version 10,11 and 12. In v4 as the doc says:

Installation assets are no longer shipped in a single monolithic package. Instead, installation assets have been split among several packages, allowing clients to opt-out of the installation of assets not applicable to their use case.

This means that users should pick the right dcgm packages based in the driver version of the host. See this compatibility table

Because of this compatibility and how NVIDIA changed how they ship the debian package, we need to change the tracks as described above.

Hope that I made it clear.

Thanks

Hi @gabrielcocenza ,

We cannot rename a track. I have created the 4 new tracks for dcgm snap as requested.

Do let me know in case of any queries.

Thanks,
Aditya