MPI GPU CLusters

[Pyrex]

I am building a rather large cluster of Jetson TK1 development boards and I have been investigating a means to implement CUDA based MPI and I found some interesting information: https://developer.nvidia.com/mpi-solutions-gpus

I started wondering how difficult it may be to use this with oclHashcat.

Does anyone else find the idea of distributing a workload across several GPU's in a cluster fascinating?

[digininja]

I've seen your picture of a few boards stacked so know you got the project started, did you get anywhere with it? If you did, how do you find the boards to work with compared to a normal desktop GPU?