Kernels implemented in NUMBA. This is an implementation of the Min-Sum algorithm for decoding of LDPC codes using a GPU. The parity matrix is not assumed, only its dimensions are needed (CUDA kernels require declaration of dimensions). To cite this repository please use:
@software{cudaLDPC, author = {Omer S. Sella}, title = {CUDA implementation of the Min-Sum decoder, code repository}, url = {https://github.com/Omer-Sella/cudaLDPC}, }