[doc-only] cuda.core: document how to install nightly (top-of-tree) builds#2175
Merged
Conversation
…uilds Adds a new section to the cuda.core installation guide explaining how to download and install the latest top-of-tree wheel artifacts published by CI on every push to main, mirroring the workflow used by numba-cuda-mlir. Closes NVIDIA#2166
Contributor
Contributor
Author
|
@leofang — when you have a moment, would you take a look? No rush. I picked you based on merged 8 docs PRs in last 6 months; happy to address feedback or re-route to someone better if you're not the right person. |
Contributor
Author
|
@mdboom — when you have a moment, would you take a look? No rush. I picked you based on merged 8 docs PRs in last 6 months; happy to address feedback or re-route to someone better if you're not the right person. |
Member
|
Please do not ping individual team members. We'll triage PRs during team sync. |
leofang
reviewed
Jun 8, 2026
leofang
added
documentation
Per @leofang's review on NVIDIA#2175: the nightly/top-of-tree install recipe is a developer concern, not end-user installation. Demote the section from a top-level Installation subsection to the first sub-subsection under Development environment (above 'Development with uv').
Contributor
Author
|
Thanks for the review, @leofang — pushed Also — sincere apology for the cold pings yesterday. I shouldn't have @-mentioned individual team members; I missed the team-triage workflow note. Won't happen again on cuda-python or other NVIDIA repos. Happy to close and resubmit without those comments if you'd prefer, but otherwise I'll just leave it for triage at your team's normal cadence. |
leofang
approved these changes
Jun 8, 2026
leofang
left a comment
leofang
left a comment
Member
There was a problem hiding this comment.
Thanks, @matysanchez! If this PR is done, let's un-draft it!
Contributor
Author
|
@leofang thanks Leo |
Member
|
/ok to test df0bb20 |
This comment has been minimized.
This comment has been minimized.
|
leofang
added a commit
to leofang/cuda-python
that referenced
this pull request
Jun 9, 2026
…rce" under "Development environment" Make "Installing from Source" consistently a subsection of "Development environment" across all three components instead of a top-level sibling, matching the intent of the developer-workflow group introduced in NVIDIA#2175. - cuda.core: drop "Installing from Source" from --- to ~~~ (now under Development environment, as a sibling of nightly / uv / pixi). - cuda.bindings: drop "Installing from Source" from --- to ~~~, and drop its "Requirements" and "Editable Install" subsections from ~~~ to ^^^ accordingly. - cuda.pathfinder: drop "Installing from Source" from --- to ~~~ in the new install.rst added earlier in this branch. Also bump SPDX copyright headers on touched existing files to 2025-2026; the new cuda_pathfinder install.rst uses 2026.
leofang
added a commit
that referenced
this pull request
Jun 10, 2026
…unify "Development environment" structure (#2187) * [doc-only] cuda.{bindings,pathfinder}: document how to install nightly (top-of-tree) builds Extends the cuda.core change in #2175 to the remaining cuda-python components named in #2166 (which also covers cuda.bindings and cuda.pathfinder). cuda_bindings: - Add a new "Development environment" section to install.rst with a "Installing the latest nightly (top-of-tree builds)" subsection, matching the cuda.core layout. Artifact pattern includes the CUDA major (cuda13*) since cuda-bindings wheels are CUDA-versioned. cuda_pathfinder: - Add a new install.rst (PyPI / Conda / Development environment with nightly subsection / Source). cuda.pathfinder previously had no standalone install page. Nightly uses the single pure-Python cuda-pathfinder-wheel artifact (no python/platform/CUDA selection needed). - Wire install.rst into the toctree in index.rst. * [doc-only] cuda.{core,bindings,pathfinder}: nest "Installing from Source" under "Development environment" Make "Installing from Source" consistently a subsection of "Development environment" across all three components instead of a top-level sibling, matching the intent of the developer-workflow group introduced in #2175. - cuda.core: drop "Installing from Source" from --- to ~~~ (now under Development environment, as a sibling of nightly / uv / pixi). - cuda.bindings: drop "Installing from Source" from --- to ~~~, and drop its "Requirements" and "Editable Install" subsections from ~~~ to ^^^ accordingly. - cuda.pathfinder: drop "Installing from Source" from --- to ~~~ in the new install.rst added earlier in this branch. Also bump SPDX copyright headers on touched existing files to 2025-2026; the new cuda_pathfinder install.rst uses 2026.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Closes #2166
What
Adds an Installing the latest nightly (top-of-tree builds) section to
cuda_core/docs/source/install.rst. The section explains how to usegh run downloadto grab the wheel artifact from the most recent successful CI run onmain, with notes for picking your Python version, target platform, and CUDA major version.Why
Per the issue, the equivalent section already exists in numba-cuda-mlir. The two repos share CI infrastructure, so the recipe transfers directly — only repo name, workflow file (
ci.yml), and artifact pattern names needed adapting.Tested
gh run list -R NVIDIA/cuda-python -w ci.yml -b main -s success -L1resolves cleanly against the live repo (returns the latest successful CI run).Verification