---
id: nvidia-cusparse-cu11
version: "11.7.5.86"
license: NVIDIA Proprietary Software
license_treatment: unclear
maintenance: abandoned
---

# nvidia-cusparse-cu11 — CUSPARSE native runtime libraries

License: unclear · Maintenance: abandoned · Downloads: 1.5M/mo

## What it is and what it does

nvidia-cusparse-cu11 is a native runtime library package that bundles NVIDIA's CUSPARSE library for CUDA 11. It provides low-level GPU acceleration for sparse matrix operations—a critical building block for machine learning frameworks and scientific computing libraries that need to perform efficient sparse linear algebra on NVIDIA GPUs.

The package is a runtime dependency, not a user-facing library; it is typically installed as a transitive dependency by higher-level frameworks like PyTorch or CuPy. It requires an NVIDIA GPU and CUDA 11 driver to function and is distributed as platform-specific wheels for Linux (x86_64 and aarch64) and Windows. However, the package has been abandoned for over 1396 days, with no recent maintenance or updates, which creates compatibility risk for users on newer CUDA or Python versions.

Use it for:

- Dependency for PyTorch or other ML frameworks configured to use CUDA 11 sparse operations.
- Scientific computing workflows requiring GPU-accelerated sparse matrix multiplication or decomposition.
- Custom CUDA applications that directly call CUSPARSE routines for sparse linear algebra.
- Legacy systems locked to CUDA 11 that cannot upgrade to newer CUDA versions.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Provides CUSPARSE native runtime libraries for CUDA 11, enabling sparse matrix operations on NVIDIA GPUs.

No—unless you are maintaining legacy code explicitly pinned to CUDA 11. The package is abandoned with no maintenance for over 1396 days, creating compatibility and security risk. Modern projects should use current NVIDIA CUDA runtime packages or rely on framework-managed CUDA dependencies instead. If you must use CUDA 11, verify that your GPU driver and CUDA toolkit match version 11.7.5.86 exactly.

## Install

pip install nvidia-cusparse-cu11
uv add nvidia-cusparse-cu11
poetry add nvidia-cusparse-cu11

## Installing nvidia-cusparse-cu11

Before you install: Medium install friction due to platform-specific wheel distribution (x86_64 Linux, aarch64, Windows). Package is abandoned—last release was over 1396 days ago with no maintenance activity, which poses risk for compatibility with newer CUDA or Python versions.

License in practice: Licensed under NVIDIA Proprietary Software with unclear treatment. Users should verify licensing terms directly with NVIDIA before deploying in production or commercial contexts.

Quickstart:

pip install nvidia-cusparse-cu11==11.7.5.86

# Import in code that uses sparse matrix operations via CUDA
import nvidia.cusparse

Requires an NVIDIA GPU with CUDA 11 support and matching driver; only available for x86_64 Linux, aarch64, or Windows platforms.

Verify before relying:

- Whether this package is still compatible with modern CUDA driver versions given its abandoned status.
- Whether the NVIDIA Proprietary Software license permits use in specific deployment contexts (commercial, cloud, etc.).
- Whether newer CUDA versions (12+) are supported or if users must pin to CUDA 11.

## Package facts

- License: NVIDIA Proprietary Software (unclear)
- Python support: supports_current
- Install friction: medium
- Maintenance: abandoned
- Downloads: 1.5M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

cuda sparse matrix operations, nvidia cusparse runtime, gpu sparse linear algebra, cusparse cuda 11, nvidia gpu acceleration, sparse tensor gpu, cuda 11 runtime libraries, gpu-runtime, cuda-legacy, sparse-linear-algebra

[View on SkillFed](https://skillfed.io/packages/nvidia-cusparse-cu11) · [View on PyPI](https://pypi.org/project/nvidia-cusparse-cu11/)