cupy/cupy

NumPy & SciPy for GPU

What it solves

CuPy solves the performance bottleneck of running large-scale numerical computations on a CPU by allowing Python users to execute NumPy and SciPy code on GPUs. It eliminates the need to rewrite existing CPU-based array code from scratch to achieve GPU acceleration.

How it works

CuPy acts as a drop-in replacement for NumPy and SciPy. It implements a compatible array library that targets NVIDIA CUDA or AMD ROCm platforms, enabling the same API calls to be executed on GPU hardware. For advanced users, it also provides low-level access to CUDA features, including RawKernels for C/C++ integration, Streams for performance optimization, and direct access to CUDA Runtime APIs.

Who it’s for

It is designed for Python developers and data scientists who use NumPy and SciPy for numerical computing and want to accelerate their workflows using NVIDIA or AMD GPUs.

Highlights

  • Drop-in compatibility with NumPy and SciPy APIs.
  • Supports both NVIDIA CUDA and AMD ROCm platforms.
  • Provides low-level CUDA C/C++ integration via RawKernels.
  • Includes integrated support for signal processing workflows (via cuSignal).

Related

  • Project
    nv-legate/cupynumeric

    A high-performance array computing library that implements the NumPy API on top of the Legate framework to scale workflows to thousands of GPUs.

  • Project
    neka-nat/cupoch

    A GPU-accelerated library for rapid 3D data processing in robotics, providing high-performance point cloud algorithms, collision avoidance, and path planning.

  • Project
    inducer/pycuda

    PyCUDA is a Python wrapper for Nvidia's CUDA API that enables high-performance parallel computation on GPUs with automatic memory management and C++ speed.

  • Project
    NVIDIA/cuda-python

    A collection of Python packages providing both high-level idiomatic and low-level binding access to NVIDIA's CUDA platform for GPU-accelerated computing.

  • Project
    rapidsai/cugraph

    A collection of GPU-accelerated graph analytics packages that enable the creation and manipulation of graphs and the execution of scalable graph algorithms.