Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/inducer/pycuda

CUDA integration for Python, plus shiny features
https://github.com/inducer/pycuda

array cuda gpu gpu-computing multidimensional-arrays pycuda python scientific-computing

Last synced: 10 days ago
JSON representation

CUDA integration for Python, plus shiny features

Awesome Lists containing this project

README

        

PyCUDA: Pythonic Access to CUDA, with Arrays and Algorithms
=============================================================

.. image:: https://gitlab.tiker.net/inducer/pycuda/badges/main/pipeline.svg
:alt: Gitlab Build Status
:target: https://gitlab.tiker.net/inducer/pycuda/commits/main
.. image:: https://badge.fury.io/py/pycuda.png
:target: https://pypi.org/project/pycuda
.. image:: https://zenodo.org/badge/1575319.svg
:alt: Zenodo DOI for latest release
:target: https://zenodo.org/badge/latestdoi/1575319

PyCUDA lets you access `Nvidia `_'s `CUDA
`_ parallel computation API from Python.
Several wrappers of the CUDA API already exist-so what's so special
about PyCUDA?

* Object cleanup tied to lifetime of objects. This idiom, often
called
`RAII `_
in C++, makes it much easier to write correct, leak- and
crash-free code. PyCUDA knows about dependencies, too, so (for
example) it won't detach from a context before all memory
allocated in it is also freed.

* Convenience. Abstractions like pycuda.driver.SourceModule and
pycuda.gpuarray.GPUArray make CUDA programming even more
convenient than with Nvidia's C-based runtime.

* Completeness. PyCUDA puts the full power of CUDA's driver API at
your disposal, if you wish. It also includes code for
interoperability with OpenGL.

* Automatic Error Checking. All CUDA errors are automatically
translated into Python exceptions.

* Speed. PyCUDA's base layer is written in C++, so all the niceties
above are virtually free.

* Helpful `Documentation `_.

Relatedly, like-minded computing goodness for `OpenCL `_
is provided by PyCUDA's sister project `PyOpenCL `_.