nvshmem4py-cu13
Python bindings for NVSHMEM
Decision gist · record as of 2026-08-14
Yes, if you are developing multi-GPU applications on NVIDIA hardware and need inter-GPU communication primitives. The package is actively maintained, has no known vulnerabilities, and offers a Pythonic entry point to NVSHMEM. However, installation requires CUDA 13 and the nvidia-nvshmem-cu13 library to be pre-installed, and the unclear license status warrants verification before use in proprietary code.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires NVIDIA CUDA 13 runtime, nvidia-nvshmem-cu13 library, and compatible GPU hardware; Linux x86_64 or aarch64 only.
- Medium install friction: requires CUDA 13 runtime, nvidia-nvshmem-cu13 system library, and Cython compilation.
- Prebuilt wheels available for Python 3.10–3.13 on x86_64 and aarch64 Linux.
License · maintenance · safety
(unclear) — License status is unclear—no SPDX identifier or raw license text is available in the package metadata. Verify licensing terms before use in proprietary or redistributed code.
last release 2026-06-11 (64 days)
0 known vulnerabilities (OSV.dev, 2026-08-14) · 552,880 downloads/mo, #6,040 on PyPI
Alternatives
Verify before relying
pip install nvshmem4py-cu13
import nvshmem4py_cu13
# Use NVSHMEM operations via the Python interface- Specific NVSHMEM API coverage and which operations are exposed through the Python interface.
- Whether this package is suitable for single-GPU or only multi-GPU workloads.
- Performance characteristics and typical use-case scale (e.g., node-local vs. multi-node).
- Compatibility with non-NVIDIA GPU frameworks or heterogeneous compute environments.
What it is and what it does
NVSHMEM4Py is a Python wrapper around NVIDIA's NVSHMEM library, which provides symmetric heap memory semantics for GPU-to-GPU communication. It allows Python developers to leverage NVSHMEM's collective operations and inter-GPU synchronization primitives without writing CUDA C/C++ code directly. The package depends on the CUDA runtime, nvidia-nvshmem-cu13 system library, cuda-python, numpy, and Cython, making it tightly coupled to NVIDIA's GPU ecosystem.
The package is actively maintained and follows NVIDIA's NVSHMEM SLA. It targets modern Python versions (3.9+) and provides prebuilt wheels for common Linux architectures. Installation requires both the CUDA 13 runtime and the underlying NVSHMEM library to be present on the system, which adds setup complexity but avoids runtime compilation for most users.
Use it for
- Implementing multi-GPU collective algorithms (all-reduce, broadcast) without writing CUDA kernels.
- Building distributed GPU applications that need low-latency inter-GPU synchronization on a single node.
- Prototyping GPU communication patterns in Python before optimizing critical paths in CUDA.
- Integrating NVSHMEM operations into existing Python-based GPU frameworks or data-parallel workflows.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you are developing multi-GPU applications on NVIDIA hardware and need inter-GPU communication primitives.
The package is actively maintained, has no known vulnerabilities, and offers a Pythonic entry point to NVSHMEM. However, installation requires CUDA 13 and the nvidia-nvshmem-cu13 library to be pre-installed, and the unclear license status warrants verification before use in proprietary code.
Install
nvshmem4py-cu13 on PyPI
Before you install
Medium install friction: requires CUDA 13 runtime, nvidia-nvshmem-cu13 system library, and Cython compilation. Prebuilt wheels available for Python 3.10–3.13 on x86_64 and aarch64 Linux. Active maintenance with recent release.
Requires NVIDIA CUDA 13 runtime, nvidia-nvshmem-cu13 library, and compatible GPU hardware; Linux x86_64 or aarch64 only.
License in practice
License status is unclear—no SPDX identifier or raw license text is available in the package metadata. Verify licensing terms before use in proprietary or redistributed code.
Quickstart
pip install nvshmem4py-cu13
import nvshmem4py_cu13
# Use NVSHMEM operations via the Python interfaceVerify before relying
- Specific NVSHMEM API coverage and which operations are exposed through the Python interface.
- Whether this package is suitable for single-GPU or only multi-GPU workloads.
- Performance characteristics and typical use-case scale (e.g., node-local vs. multi-node).
- Compatibility with non-NVIDIA GPU frameworks or heterogeneous compute environments.
Package facts
| License | Not declared unclear |
| Python support | Supports the current Python release >=3.9 |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 6 packagesnvidia-nvshmem-cu13cuda-pythoncuda.corecuda.pathfindernumpyCython |
| Maintenance | Actively maintained 64 days since the last release |
| First released | |
| Downloads | 552,880 / month, #6,040 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
Evidence: nvshmem4py_cu13-0.3.1-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_34_aarch64.whl; nvshmem4py_cu13-0.3.1-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_34_x86_64.whl; nvshmem4py_cu13-0.3.1-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_34_aarch64.whl; nvshmem4py_cu13-0.3.1-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_34_x86_64.whl; nvshmem4py_cu13-0.3.1-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_34_aarch64.whl; nvshmem4py_cu13-0.3.1-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_34_x86_64.whl; nvshmem4py_cu13-0.3.1-cp313-cp313-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_34_aarch64.whl; nvshmem4py_cu13-0.3.1-cp313-cp313-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_34_x86_64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “NVSHMEM python bindings”
- nvshmem4py-cu13Provides a Python interface to NVSHMEM, enabling GPU-accelerated…
- nvidia-nvshmem-cu13NVSHMEM provides a global address space for GPU cluster…
- nvidia-nvshmem-cu12NVSHMEM provides a global address space for GPU cluster…
Give your agent the search over MCP, or paste the wish link into any chat.
More Distributed Computing packages
gRPC Python is an HTTP/2-based RPC framework that enables you to define and call remote procedures across network boundaries using protocol buffers for serialization.
Install it if you need RPC communication in a distributed system or are integrating with existing gRPC services.
execnet lets you spawn and communicate with Python interpreters across local processes, remote hosts, and different platforms, using a simple API for task distribution and inter-process messaging.
However, the aging maintenance status (275 days since last release) means you should verify it meets your concurrency and performance needs before committing to a…
Cloudpickle extends Python's standard pickle module to serialize lambda functions, interactively-defined functions and classes, and other constructs that the default pickle cannot handle, making it suitable for cluster computing and remote code execution.
Install it if you need to serialize lambda functions, interactively-defined code, or non-standard Python constructs for cluster computing or distributed execution.
Provides a unified, open()-compatible Python API for streaming large files from remote storage (S3, GCS, Azure, HDFS, SFTP, HTTP) and local filesystems, with transparent compression support.
Install it if you work with large files on cloud storage or remote systems and want to avoid writing boilerplate around multiple SDKs.
Portalocker provides cross-platform file locking with support for exclusive and shared locks, plus Redis-based distributed locks and process-aware PID file locking.
Install it if you need file or process coordination; the optional extras (pywin32, redis) are only required for specific lock types.
Ray is a distributed computing framework that scales Python applications from a single machine to multi-node clusters, providing abstractions for parallel tasks, stateful actors, and shared objects.
See also nccl4py · nvidia-nvshmem-cu12 · nvidia-nvshmem-cu13 · cuda-python · pycuda · nvidia-nccl-cu13 · custatevec-cu13 · nvidia-cublas-cu11 · nvidia-cusolver-cu11 · nvidia-cusparse-cu12