Hello. I have seen a couple of threads about the possibility of using rapids in a jetson. In some of then it is said that there is no docker ARM image and it others it says that it might be possible. Does anyone know if there is a straight away way to install it in an AGX Xavier?
Please check this GitHub for Rapids on Xavier:
I think that that repo is deprecated because of a comment in the repo it has been forked from. Indeed, the comment is from JasonAtNvidia:
this repo is not up to date. All necessary things required to build RAPIDS for aarch64 have been merged into the RAPIDS project itself. Follow the instructions to build RAPIDS from the actual github.com/rapidsai/* repositories. cuStrings doesn’t even exist any longer in modern RAPIDS.
@awthomp consider deleting this repository, hiding it from public, or adding a not maintained into the README file for this repo.
Should I install it as explained in cudf/CONTRIBUTING.md at branch-0.19 · rapidsai/cudf · GitHub Build from Source?
Sorry that we don’t have too much experience on the RAPIDS.
Would you mind to give it a try and share the result or error (if any) with us?
I have tried a couple of things. I leave them here in case anybody wants more information:
- The link you passed until I realized that it was outdated and now there were non-compatible packages.
- This repository: GitHub - gdaisukesuzuki/PKGBUILD_Rapids: Building NVIDIA's RAPIDS (cuDF, cuML...) in Arch Linux
It might help bu,t I as I have not used the PKGBUILD format before, I am having trouble installing what I need to use that format.
Hi, do you have any updates? Were you successful eventually? I have a jetson xavier nx. I would like to install cuML and run some codes on it. Any help would be appreciated.
No, I didn’t achieve it. I had to change the approach to my problem.