René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Make `cuda-bindings` buildable against CTK wheels?","articleBody":"The meaning of the title is that `pip install .` from source can work without a local CTK environment (no need to set `CUDA_PATH` and all that). Everything is purely pip-wheel based.\n\nThis is actually doable, I just don't believe there is a serious use case because we really want users to not build from source, nor do I want to spend my time on enabling it, but I can be convinced. Anyway it's worth writing this up somewhere in the public.\n\nThe whole thing relies on the fact that building bindings just need a host compiler; no NVCC is needed. The steps is as follows:\n- We [list needed CTK wheels](https://github.com/NVIDIA/nvmath-python/blob/1cbc60be436fe98e8a08f38e9867bc6fba90f007/pyproject.toml#L8-L10) in the `[build-system.requires]` session. They will provide the needed headers.\n- Since `setuptools` does not provide a way for us to programmatically query the path to the isolated build environment at build time, [we search the potential paths](https://github.com/NVIDIA/nvmath-python/blob/1cbc60be436fe98e8a08f38e9867bc6fba90f007/builder/utils.py#L12-L41) based on the assumption that `sys.path` must be updated for `setuptools` to find the build env.\n- We [pass `\u003cbuild env path>` to `setuptools`](https://github.com/NVIDIA/nvmath-python/blob/1cbc60be436fe98e8a08f38e9867bc6fba90f007/builder/utils.py#L118) so that the CUDA headers can be found by gcc/msvc.\n\nI enabled this for nvmath-python and it did worked like a charm, but the situation was a lot simpler than driver/runtime modules that we have here.","author":{"url":"https://github.com/leofang","@type":"Person","name":"leofang"},"datePublished":"2025-06-07T01:32:45Z","url":"https://github.com/NVIDIA/cuda-python/issues/692"}