Hi, Are there any recommendations for running embarrassingly parallel problems across multiple windows machines?
I have to do the same piece of work on multiple multiple GB files - which takes hours on a single machine spun out across the cores with parallel map. I'm lucky enough to have access to several windows machines and also a couple of debian ones and was wondering what the best way to split the job across them I've copied the source data across all the machines. Is there some equivalent of pmap that anyone has running across multiple machines (e.g. with ssh or mpi)? or should I just write my own start up and wind down scripts and in that split the work into chunks for each machine communicating via ssh or whatever? Set up linux virtual machines and then use the current addprocs with extra machines? Thanks for any thoughts.
