Sinkhorn-Knopp-Parallel-CUDA
This is a parallel implementation of Sinkhorn-Knopp algorithm created for Parallel Computing class in Sabanci University. The input is taken as compressed row and column representation. The results are discussed in the report which can also be found in the repo.
I do not generally share homeworks, class projects etc. but I realized that there are not much CUDA based Sinkhorn-Knopp that can be found easily. The code is not at its best form and there are more optimizations that could be done on this especially regarding usage of shared memory. So keep in mind this is not "fully optimized".