Install this version:
emerge -a =dev-python/cuda-python-13.2.0
If this version is masked, you can unmask it using the autounmask tool or standard emerge options:
autounmask =dev-python/cuda-python-13.2.0
Or alternatively:
emerge --autounmask-write -a =dev-python/cuda-python-13.2.0
| Version | EAPI | Keywords | Slot |
|---|---|---|---|
| 13.2.0 | 8 | ~amd64 | 0 |
# Copyright 1999-2026 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2
EAPI=8
DISTUTILS_USE_PEP517=no
PYTHON_COMPAT=( python3_{11..14} )
inherit distutils-r1
MY_WHEEL="${PN//-/_}-${PV}-py3-none-any.whl"
DESCRIPTION="Metapackage pulling cuda-bindings and cuda-pathfinder"
HOMEPAGE="
https://github.com/NVIDIA/cuda-python
https://nvidia.github.io/cuda-python/
https://pypi.org/project/cuda-python/
"
SRC_URI="
https://files.pythonhosted.org/packages/4a/da/b4dbe129f941afe1c24a09ba53521b78875626763d96414798a74763282f/${MY_WHEEL}
"
S="${WORKDIR}"
LICENSE="NVIDIA-CUDA"
SLOT="0"
KEYWORDS="~amd64"
# NVIDIA-CUDA is an EULA license — distfile must not be mirrored,
# resulting binpkgs must not be redistributed.
RESTRICT="bindist mirror"
# Trivial meta-wheel — empty package whose entire purpose is to pull
# cuda-bindings + cuda-pathfinder via runtime deps. Source is just
# pyproject.toml in the cuda-python/ subdir of NVIDIA/cuda-python with
# no Python code; packaging the wheel directly is exactly equivalent
# and avoids carrying the monorepo tarball for an empty payload.
RDEPEND="
~dev-python/cuda-bindings-${PV}[${PYTHON_USEDEP}]
dev-python/cuda-pathfinder[${PYTHON_USEDEP}]
"
src_unpack() {
mkdir -p "${S}/wheel" || die
cp "${DISTDIR}/${MY_WHEEL}" "${S}/wheel/" || die
}
src_install() {
python_foreach_impl install_wheel
}
install_wheel() {
${EPYTHON} -m installer --destdir="${D}" "${S}/wheel/${MY_WHEEL}" || die
}
~dev-python/cuda-bindings-13.2.0[] dev-python/cuda-pathfinder[]