Installation

Installing Conda Packages

cuPyNumeric supports the same platforms as Legate.

cuPyNumeric is available from conda on the legate channel.

Note

conda version >= 24.1 required

# with a new environment
$ conda create -n myenv -c conda-forge -c legate cupynumeric

# =========== OR =========== #

# into an existing environment
$ conda install -c conda-forge -c legate cupynumeric

Important

Packages are offered for Linux (x86_64 and aarch64) and macOS (aarch64, pip wheels only), supporting Python versions 3.11 to 3.13. Windows is only supported through WSL.

Nightly top-of-tree builds of cuPyNumeric are available under the “experimental” label:

Refer to the nightly docs when using these.

Important

These builds are only lightly validated. Use them at your own risk.

Installing PyPI Packages

cuPyNumeric is also available from PyPI. To install, run the following command:

# into existing environment
$ pip install nvidia-cupynumeric

# =========== OR =========== #

# into new environment
$ python -m venv myenv
$ source myenv/bin/activate
$ pip install nvidia-cupynumeric

This will install the latest version of cuPyNumeric and the corresponding version of Legate.

The cuPyNumeric package on PyPI is multi-node and multi-rank capable. Please check Legate documentation to find more details about running on multiple nodes.

Important

Packages are offered for Linux (x86_64 and aarch64) and macOS (aarch64, pip wheels only), supporting Python versions 3.11 to 3.13. Windows is only supported through WSL.

Verify your Installation

You can verify the installation by running one of the examples.

For instance:

$ legate examples/black_scholes.py
Running black scholes on 10K options...
Elapsed Time: 129.017 ms

Conda and GPU / CPU Variants

conda automatically installs the right variant for the system: * CPU variant if no NVIDIA GPU is detected * GPU variant if an NVIDIA GPU is detected

To override this behavior and force install a version with GPU support, use the following (with the desired CUDA version):

$ CONDA_OVERRIDE_CUDA="12.2" conda install -c conda-forge -c legate cupynumeric

Building from source

See Building from source for instructions on building cuPyNumeric manually.

Licenses

This project will download and install additional third-party open source software projects at install time. Review the license terms of these open source projects before use.

For license information regarding projects bundled directly, see Third-party notices.