Ecosyste.ms: Awesome

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

https://github.com/cupy/cupy

NumPy & SciPy for GPU
https://github.com/cupy/cupy

cublas cuda cudnn cupy curand cusolver cusparse cusparselt cutensor gpu nccl numpy nvrtc nvtx python rocm scipy tensor

Last synced: 18 days ago
JSON representation

NumPy & SciPy for GPU

Lists

README 

        



# CuPy : NumPy & SciPy for GPU



[![pypi](https://img.shields.io/pypi/v/cupy)](https://pypi.python.org/pypi/cupy)

[![Conda](https://img.shields.io/conda/vn/conda-forge/cupy)](https://anaconda.org/conda-forge/cupy)

[![GitHub license](https://img.shields.io/github/license/cupy/cupy)](https://github.com/cupy/cupy)

[![Gitter](https://badges.gitter.im/cupy/community.svg)](https://gitter.im/cupy/community)

[![Twitter](https://img.shields.io/twitter/follow/CuPy_Team?label=%40CuPy_Team)](https://twitter.com/CuPy_Team)



[**Website**](https://cupy.dev/)

| [**Install**](https://docs.cupy.dev/en/stable/install.html)

| [**Tutorial**](https://docs.cupy.dev/en/stable/user_guide/basic.html)

| [**Examples**](https://github.com/cupy/cupy/tree/main/examples)

| [**Documentation**](https://docs.cupy.dev/en/stable/)

| [**API Reference**](https://docs.cupy.dev/en/stable/reference/)

| [**Forum**](https://groups.google.com/forum/#!forum/cupy)



CuPy is a NumPy/SciPy-compatible array library for GPU-accelerated computing with Python.

CuPy acts as a [drop-in replacement](https://docs.cupy.dev/en/stable/reference/comparison.html) to run existing NumPy/SciPy code on NVIDIA CUDA or AMD ROCm platforms.



```py

>>> import cupy as cp

>>> x = cp.arange(6).reshape(2, 3).astype('f')

>>> x

array([[ 0.,  1.,  2.],

       [ 3.,  4.,  5.]], dtype=float32)

>>> x.sum(axis=1)

array([  3.,  12.], dtype=float32)

```



CuPy also provides access to low-level CUDA features.

You can pass `ndarray` to existing CUDA C/C++ programs via [RawKernels](https://docs.cupy.dev/en/stable/user_guide/kernel.html#raw-kernels), use [Streams](https://docs.cupy.dev/en/stable/reference/cuda.html) for performance, or even call [CUDA Runtime APIs](https://docs.cupy.dev/en/stable/reference/cuda.html#runtime-api) directly.



## Installation



### Pip



Binary packages (wheels) are available for Linux and Windows on [PyPI](https://pypi.org/org/cupy/).

Choose the right package for your platform.



| Platform              | Architecture      | Command                                                       |

| --------------------- | ----------------- | ------------------------------------------------------------- |

| CUDA 11.x (11.2+)     | x86_64 / aarch64  | `pip install cupy-cuda11x`                                    |

| CUDA 12.x             | x86_64 / aarch64  | `pip install cupy-cuda12x`                                    |

| ROCm 4.3 (*[experimental](https://docs.cupy.dev/en/latest/install.html#using-cupy-on-amd-gpu-experimental)*)          | x86_64            | `pip install cupy-rocm-4-3`                                   |

| ROCm 5.0 (*[experimental](https://docs.cupy.dev/en/latest/install.html#using-cupy-on-amd-gpu-experimental)*)          | x86_64            | `pip install cupy-rocm-5-0`                                   |



> [!NOTE]\

> To install pre-releases, append `--pre -U -f https://pip.cupy.dev/pre` (e.g., `pip install cupy-cuda11x --pre -U -f https://pip.cupy.dev/pre`).



### Conda



Binary packages are also available for Linux and Windows on [Conda-Forge](https://anaconda.org/conda-forge/cupy).



| Platform              | Architecture                | Command                                                       |

| --------------------- | --------------------------- | ------------------------------------------------------------- |

| CUDA                  | x86_64 / aarch64 / ppc64le  | `conda install -c conda-forge cupy`                           |



If you need a slim installation (without also getting CUDA dependencies installed), you can do `conda install -c conda-forge cupy-core`.



If you need to use a particular CUDA version (say 12.0), you can use the `cuda-version` metapackage to select the version, e.g. `conda install -c conda-forge cupy cuda-version=12.0`.



> [!NOTE]\

> If you encounter any problem with CuPy installed from `conda-forge`, please feel free to report to [cupy-feedstock](https://github.com/conda-forge/cupy-feedstock/issues), and we will help investigate if it is just a packaging issue in `conda-forge`'s recipe or a real issue in CuPy.



### Docker



Use [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/overview.html) to run [CuPy container images](https://hub.docker.com/r/cupy/cupy).



```

$ docker run --gpus all -it cupy/cupy

```



## Resources



- [Installation Guide](https://docs.cupy.dev/en/stable/install.html) - instructions on building from source

- [Release Notes](https://github.com/cupy/cupy/releases)

- [Projects using CuPy](https://github.com/cupy/cupy/wiki/Projects-using-CuPy)

- [Contribution Guide](https://docs.cupy.dev/en/stable/contribution.html)



## License



MIT License (see `LICENSE` file).



CuPy is designed based on NumPy's API and SciPy's API (see `docs/source/license.rst` file).



CuPy is being developed and maintained by [Preferred Networks](https://www.preferred.jp/en/) and [community contributors](https://github.com/cupy/cupy/graphs/contributors).



## Reference



Ryosuke Okuta, Yuya Unno, Daisuke Nishino, Shohei Hido and Crissman Loomis.

**CuPy: A NumPy-Compatible Library for NVIDIA GPU Calculations.**

*Proceedings of Workshop on Machine Learning Systems (LearningSys) in The Thirty-first Annual Conference on Neural Information Processing Systems (NIPS)*, (2017).

[[PDF](http://learningsys.org/nips17/assets/papers/paper_16.pdf)]



```bibtex

@inproceedings{cupy_learningsys2017,

  author       = "Okuta, Ryosuke and Unno, Yuya and Nishino, Daisuke and Hido, Shohei and Loomis, Crissman",

  title        = "CuPy: A NumPy-Compatible Library for NVIDIA GPU Calculations",

  booktitle    = "Proceedings of Workshop on Machine Learning Systems (LearningSys) in The Thirty-first Annual Conference on Neural Information Processing Systems (NIPS)",

  year         = "2017",

  url          = "http://learningsys.org/nips17/assets/papers/paper_16.pdf"

}

```