is a boilerplate project to document CUDA-bindings for Numpy using Swig. pyCUREVERSE does not claim
- to have any application in real world scenarios
- to be an efficient reversal algorithm of arrays on CUDA-enabled devices
Nevertheless, you can use pyCUDAREVERSE
- to bind any CUDA-kernel to numpy arrays (fuck yeah CUFFT, CUBLAS, ...)
- to have fun with GPUs
- to visualize the stuff (matplotlib, mayavi) you have calculate with CUDA
Please note, there is pycuda. pyCUREVERSE aims to provide bindings on the lowest level possible.
git clone https://github.com/gravitino/pyCUREVERSE.git
cd pyCUREVERSE
vim Makefile (make sure the paths to CUDA are OK)
make rename NEWLIBNAME="myAwesomeLib"
make
python example.py
Have fun.