Build and test with CUDA 13.0.0

[jameslamb]

Contributes to rapidsai/build-planning#208

• uses CUDA 13.0.0 to build and test
• adds CUDA 13 devcontainers
• moves some dependency pins:
  • cupy: >=13.6.0
• temporarily skips installing cudf, dask-cuda, and dask-cudf in CUDA 13 environments (those projects do not have CUDA 13 packages yet)

Contributes to rapidsai/build-planning#68

• updates to CUDA 13 dependencies in fallback entries in dependencies.yaml matrices (i.e., the ones that get written to pyproject.toml in source control)

Notes for Reviewers

This switches GitHub Actions workflows to the cuda13.0 branch from here: rapidsai/shared-workflows#413

A future round of PRs will revert that back to branch-25.10, once all of RAPIDS supports CUDA 13.

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[jameslamb]

/ok to test

[jameslamb]

/ok to test

[jameslamb]

Problem 1: tests and devcontainers require CUDA 13 cudf packages, which don't exist yet

error    libmamba Could not solve for environment specs
    The following packages are incompatible
    ├─ cuda-version =13.0 * is requested and can be installed;
    └─ cudf =25.10,>=0.0.0a0 * is not installable because it requires
       └─ cuda-version >=12,<13.0a0 *, which conflicts with any installable versions previously reported.
critical libmamba Could not solve for environment specs

(conda-python-distributed-tests link)

ERROR: Could not find a version that satisfies the requirement cudf-cu13==25.10.*,>=0.0.0a0 (from versions: 0.0.0a0)

ERROR: No matching distribution found for cudf-cu13==25.10.*,>=0.0.0a0

(pip devcontainers link)

[pentschev]

Problem 1: tests and devcontainers require CUDA 13 cudf packages, which don't exist yet

We can probably temporarily remove it from dependencies.yaml for CUDA 13 only, the tests are supposed to run and skip cuDF tests if it's not installed, if anything fails (because we're missing the proper skips) then it may be a bug that we can fix.

[jameslamb]

We can probably temporarily remove it from dependencies.yaml for CUDA 13 only, the tests are supposed to run and skip cuDF tests if it's not installed

Ah ok, I'll try that!

I support that, and I'll add a note to rapidsai/build-planning#208 to ensure we don't consider CUDA 13 support in RAPIDS "done" until that has been reverted and the tests that need cudf here are running with CUDA 13.

[jameslamb]

/ok to test

[jameslamb]

Alright looks like that's working!

Seeing lots of passing tests on CUDA 13 and the pip CUDA 13 devcontainers passing 🎉 (build link)

Looks like just the conda devcontainers are failing... I think dask-cuda and dask-cudf should similarly be excluded for now (they make it into devcontainers via the dev: dependency list in dependencies.yaml):

Could not solve for environment specs
The following packages are incompatible
├─ cuda-version =13.0 * is requested and can be installed;
└─ dask-cudf =25.10,>=0.0.0a0 * is not installable because it requires
   └─ cuda-version >=12,<13.0a0 *, which conflicts with any installable versions previously reported.

(conda devcontainers build link)

I've pushed 9f6d991 proposing that.

[jameslamb]

/ok to test

[pentschev]

LGTM, thanks @jameslamb !

[jameslamb]

/merge