Registry / ai-ml / nvidia-cuda-cccl-cu12

nvidia-cuda-cccl-cu12

library12.9.27pypiunverified

The `nvidia-cuda-cccl-cu12` package provides the CUDA C++ Core Library (CCCL) headers and potentially compiled artifacts for CUDA Toolkit 12.x. CCCL is a collection of essential C++ utilities and primitives (like Thrust, CUB, libcudacxx) used for high-performance CUDA programming. This package is primarily a build and runtime dependency for other Python libraries that leverage specific CUDA Toolkit versions and C++ features, rather than offering direct Python APIs for end-user applications.

ai-ml

pip install nvidia-cuda-cccl-cu12

homepage ↗

INSTALL

IMPORT

SIG · NVIDIA-CUDA-CCCL-C

nvidia-cuda-cccl-cu12

ai-mlenv12.9.27

Install

1.9s avg

Import

—

Disk

33MB

Pass rate

5/ 10

Env Coverage5 / 10

glibc

3.9–3.13

musl

3.9–3.13

Install & Compatibility

Where this runs

tested against v12.9.27 · pip install

no network on importno background threads

Install × environment matrix

Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux

musl

py 3.10–3.910 runs

build_error

glibc

py 3.10–3.910 runs

installs and imports cleanly · install 1.9s · import 0.000s · 35MB

33MB installed

● package 33MB

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

N/A

✓ Not applicable; this package is not for direct Python import.

`nvidia-cuda-cccl-cu12` is a low-level C++ library component of the CUDA Toolkit, packaged as a wheel. It serves as a build and runtime dependency for other Python libraries (e.g., deep learning frameworks) that are compiled against a specific CUDA Toolkit version. It is not intended for direct Python import or usage by end-user applications.

The `nvidia-cuda-cccl-cu12` package does not offer direct Python APIs for application development. Its installation primarily provides the underlying C++ core libraries for other Python packages that build against or rely on CUDA Toolkit 12.x. Therefore, there is no functional Python code to 'use' this package directly. The example demonstrates that it's not designed for direct Python import and suggests how one might confirm its presence.

# This package provides C++ libraries and headers, not direct Python APIs. # Attempting to import it directly will likely fail or yield empty results. try: # This import is illustrative and not expected to provide useful symbols. # The actual C++ components are linked at build-time by other libraries. import nvidia_cuda_cccl_cu12 # Hypothetical; this package is generally not importable as a module print("Warning: Direct import of nvidia_cuda_cccl_cu12 succeeded but is not intended for usage.") print("This package serves as a backend dependency for other CUDA-enabled libraries.") except ImportError: print("As expected, nvidia_cuda_cccl_cu12 is not directly importable as a Python module for application code.") print("Its purpose is to provide C++ headers and libraries for other packages that build against CUDA.") # To verify installation, you would typically use pip show: # import subprocess # subprocess.run(['pip', 'show', 'nvidia-cuda-cccl-cu12'])

Debug

Known issues

breakingABI incompatibility due to CUDA version mismatch.

fix

Ensure that the `nvidia-cuda-cccl-cuXX` version (e.g., `cu12`) precisely matches the CUDA Toolkit version used by other Python libraries (like PyTorch, TensorFlow) and your GPU driver. Mismatches can lead to runtime errors, segmentation faults, or unexpected behavior, as the C++ ABI can change between CUDA versions.

affects: All versions

gotchaNot for direct Python API usage by end-user applications.

fix

Do not expect to import symbols or call functions from `nvidia-cuda-cccl-cu12` in your Python code. Its purpose is to provide underlying C++ components during the build and runtime of other CUDA-dependent Python libraries. If you are looking for Python bindings to CUDA features, explore libraries like `pycuda`, `numba`, or deep learning frameworks such as PyTorch or TensorFlow.

affects: All versions

gotchaTight coupling with the NVIDIA CUDA Toolkit.

fix

This package is a component of the larger NVIDIA CUDA Toolkit. It is crucial to have a compatible NVIDIA GPU and a correctly installed CUDA Toolkit on your system. This PyPI package primarily facilitates access to the C++ core library components for Python environments, assuming the underlying CUDA infrastructure is present and compatible.

affects: All versions

Upgrade

Version history

12.9.27latest on PyPI · released May 1, 2025

Audit

Dependencies

No dependency data recorded yet.

Agent activity

5 hits · last 30 days

node

Resources

homepagehttps://developer.nvidia.com/cuda-toolkit ↗