Cryptanalysis of algorithm using CUDA. Need info

I’m going to write parallel program for cryptanalysis of certain encryption algorithm. I plan to use differential cryptanalysis (or linear). The only question is which algorithm to use. DES - not suitable. hash functions - as well.
Maybe someone somewhere met some articles, papers or other information about cryptanalysis of algorithm that is easy to parallelize. if someone could advise something. there is lot of info in web, but mostly everywhere are used hash functions or DES with reduced number of rounds.
I would appreciate any information (or advice).

enyone? something?