nvidia-cuda-cupti-cu11

What it is and what it does

This package bundles NVIDIA's CUDA Unified Profiling Tools Interface (CUPTI) runtime libraries, which enable third-party profiling and monitoring tools to access GPU performance metrics and trace data. It is a low-level dependency typically installed automatically by higher-level GPU frameworks rather than used directly.

The package targets developers working with CUDA 11 who need profiling capabilities for GPU applications. It supports Python 3.5 through 3.11 and is distributed as platform-specific wheels for Windows and Linux (x86_64 and ARM64). However, the package has been abandoned since its latest release on 2022-10-03, making it a legacy component that may not support newer CUDA versions or Python releases.

Use it for:

• Enable GPU profiling in CUDA 11-based machine learning frameworks that depend on CUPTI for performance monitoring
• Support third-party profiling tools that require NVIDIA's GPU profiling APIs on Linux or Windows systems
• Provide runtime libraries for legacy GPU applications built against CUDA 11 that need performance analysis

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides NVIDIA CUDA profiling runtime libraries (CUPTI) for GPU performance analysis and monitoring on systems with CUDA 11 support.

No, unless you are maintaining a legacy CUDA 11 application that explicitly requires this package. The abandoned status and proprietary licensing make it a poor choice for new projects. Check whether your framework or profiling tool actually needs this specific package version, or use a maintained NVIDIA CUDA package instead.

Install

nvidia-cuda-cupti-cu11 on PyPI

pip install nvidia-cuda-cupti-cu11

uv add nvidia-cuda-cupti-cu11

poetry add nvidia-cuda-cupti-cu11

Installing nvidia-cuda-cupti-cu11

Before you install

Package is abandoned with no maintenance since its latest release on 2022-10-03. Medium install friction due to platform-specific wheel distribution across Windows and Linux architectures. No runtime dependencies to manage.

License in practice

Licensed under NVIDIA Proprietary Software with unclear treatment. Review NVIDIA's licensing terms before use in commercial or redistributed contexts.

Quickstart

pip install nvidia-cuda-cupti-cu11==11.8.87

import nvidia.cuda_cupti_cu11

Requires CUDA 11 toolkit installed on system; platform-specific wheels for Windows, Linux x86_64, or Linux ARM64 only.

Verify before relying

• Whether this package works with CUDA versions beyond 11.8.87 or if newer NVIDIA CUDA packages should be used instead
• Compatibility with Python versions beyond 3.11 given the abandoned maintenance status
• Whether direct use is necessary or if a higher-level framework handles installation automatically

Package facts

Evidence: nvidia_cuda_cupti_cu11-11.8.87-py3-none-manylinux1_x86_64.whl; nvidia_cuda_cupti_cu11-11.8.87-py3-none-manylinux2014_aarch64.whl; nvidia_cuda_cupti_cu11-11.8.87-py3-none-manylinux2014_x86_64.whl; nvidia_cuda_cupti_cu11-11.8.87-py3-none-win_amd64.whl

Keywords: cuda, nvidia, runtime, machine learning, deep learning

Tags

More Software Development packages