Python-Cuda-Extension

1. Introduction

This is a python-cuda extension which aim to provide tensor acceleration. This extension now support a number of basic operations and more functions will be added to this project in the future.

Please remember that THIS PROJECT IS COMPLETELY DIFFERENT FROM THE NVIDIA PYCUDA LIBRARY.

2. Build Python-Cuda Extension

2.1 Environment

Please make sure that your cuda, cmake, python (>=3.9), and numpy have been installed correctly before running the scripts below.

2.2 Build on Windows

git clone https://github.com/Att100/Python-Cuda-Extension.git
cd Python-Cuda-Extension
git clone https://github.com/pybind/pybind11.git
./scripts/build_pycuda_win.cmd

2.3 Build on Linux

git clone https://github.com/Att100/Python-Cuda-Extension.git
cd Python-Cuda-Extension
git clone https://github.com/pybind/pybind11.git
sh ./scripts/build_pycuda_linux.sh

2.4 Run Tests

python run_tests.py

# output -->
Test: constructor/numpy => [status: Passed, time: 0.00441 sec]
Test: matmul => [status: Passed, time: 0.00103 sec]
Test: transpose => [status: Passed, time: 0.00064 sec]
All tests passed!

3. Usage

Please reference the documents in doc for guidence.

References

• pybind11