This repo contains a parallelized groestl hash implementation for NVidia GPU

The Q transformation can be parallelized(since it is not chained) so first we transform the whole input message using the Q transformation only and then we use 8 parallel threads to apply the P transformation. The input can of any size N.

GPU Version: hash.cu  (Cuda-fied) 
Desktop version : hash.c (used for reference)

This builds on top of implementation from -> from https://www.groestl.info