From 6479b7facb1fbf0e339de53400375ada9bc45ba5 Mon Sep 17 00:00:00 2001 From: matysanchez Date: Sat, 6 Jun 2026 19:02:56 -0500 Subject: [PATCH 1/2] [doc-only] cuda.core: document how to install nightly (top-of-tree) builds 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 #2166 --- cuda_core/docs/source/install.rst | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) diff --git a/cuda_core/docs/source/install.rst b/cuda_core/docs/source/install.rst index 33a46a8c84e..76e1b923bc1 100644 --- a/cuda_core/docs/source/install.rst +++ b/cuda_core/docs/source/install.rst @@ -71,6 +71,32 @@ and likewise use ``cuda-version=13`` for CUDA 13. Note that to use ``cuda.core`` with nvJitLink installed from conda-forge requires ``cuda.bindings`` 12.8.0+. +Installing the latest nightly (top-of-tree builds) +-------------------------------------------------- + +These are useful for users looking to test new features or bug fixes prior to +their inclusion in a release. + +CI publishes wheels as GitHub Actions artifacts on every push to ``main``. To +obtain the most recent build, use the following commands: + +.. code-block:: console + + $ # Find the latest successful CI run on main: + $ RUN_ID=$(gh run list -R NVIDIA/cuda-python -w ci.yml -b main -s success -L1 --json databaseId -q '.[0].databaseId') + + $ # Download the wheel (pick your Python version and platform): + $ gh run download "$RUN_ID" -R NVIDIA/cuda-python -p "cuda-core-python312-linux-64-*" + + $ # Install the downloaded wheel: + $ pip install cuda-core-python312-linux-64-*/cuda_core*.whl[cu13] + +Replace ``python312`` with your Python version (e.g. ``python310``, ``python311``, +``python313``, ``python314``, ``python314t``). For aarch64, replace ``linux-64`` +with ``linux-aarch64``; for Windows, use ``win-64``. Replace ``cu13`` with +``cu12`` for CUDA 12.x environments. + + Development environment ----------------------- From c4319d876b1c4486042a8674927f9f6abd078142 Mon Sep 17 00:00:00 2001 From: matysanchez Date: Mon, 8 Jun 2026 11:35:56 -0500 Subject: [PATCH 2/2] [doc-only] cuda.core: move nightly install under Development environment Per @leofang's review on #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'). --- cuda_core/docs/source/install.rst | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/cuda_core/docs/source/install.rst b/cuda_core/docs/source/install.rst index 76e1b923bc1..0aa0faf58e9 100644 --- a/cuda_core/docs/source/install.rst +++ b/cuda_core/docs/source/install.rst @@ -71,8 +71,14 @@ and likewise use ``cuda-version=13`` for CUDA 13. Note that to use ``cuda.core`` with nvJitLink installed from conda-forge requires ``cuda.bindings`` 12.8.0+. +Development environment +----------------------- + +The sections above cover end-user installation. The section below focuses on +a repeatable *development* workflow (editable installs and running tests). + Installing the latest nightly (top-of-tree builds) --------------------------------------------------- +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ These are useful for users looking to test new features or bug fixes prior to their inclusion in a release. @@ -96,13 +102,6 @@ Replace ``python312`` with your Python version (e.g. ``python310``, ``python311` with ``linux-aarch64``; for Windows, use ``win-64``. Replace ``cu13`` with ``cu12`` for CUDA 12.x environments. - -Development environment ------------------------ - -The sections above cover end-user installation. The section below focuses on -a repeatable *development* workflow (editable installs and running tests). - Development with uv ~~~~~~~~~~~~~~~~~~~