Installation

Runtime Requirements

cuda.bindings supports the same platforms as CUDA. Runtime dependencies are:

  • Driver: Linux (450.80.02 or later) Windows (456.38 or later)

  • CUDA Toolkit 12.x

Note

Only the NVRTC and nvJitLink redistributable components are required from the CUDA Toolkit, which can be obtained via PyPI, Conda, or local installers (as described in the CUDA Toolkit Windows and Linux Installation Guides).

Starting from v12.8.0, cuda-python becomes a meta package which currently depends only on cuda-bindings; in the future more sub-packages will be added to cuda-python. In the instructions below, we still use cuda-python as example to serve existing users, but everything is applicable to cuda-bindings as well.

Installing from PyPI

$ pip install -U cuda-python

Install all optional dependencies with:

pip install -U cuda-python[all]

Where the optional dependencies are:

  • nvidia-cuda-nvrtc-cu12 (Provides NVRTC shared library)

  • nvidia-nvjitlink-cu12>=12.3 (Provides nvJitLink shared library)

Installing from Conda

$ conda install -c conda-forge cuda-python

Installing from Source

Requirements

  • CUDA Toolkit headers[1]

Source builds require that the provided CUDA headers are of the same major.minor version as the cuda.bindings you’re trying to build. Despite this requirement, note that the minor version compatibility is still maintained. Use the CUDA_HOME (or CUDA_PATH) environment variable to specify the location of your headers. For example, if your headers are located in /usr/local/cuda/include, then you should set CUDA_HOME with:

$ export CUDA_HOME=/usr/local/cuda

See Environment Variables for a description of other build-time environment variables.

Note

Only cydriver, cyruntime and cynvrtc are impacted by the header requirement.

Editable Install

You can use

$ pip install -v -e .

to install the module as editable in your current Python environment (e.g. for testing of porting other libraries to use the binding).