Although spark is nice, I'm also looking forward to mpi/orted integration with hadoop... "Performance: Launches ~1000x faster, runs ~10x faster" "Launch scali

Although spark is nice, I'm also looking forward to mpi/orted integration with hadoop... "Performance: Launches ~1000x faster, runs ~10x faster" "Launch scaling: Hadoop (~N), MR+ (~logN)" "Wireup: Hadoop (~N2), MR+ (~logN)" http://slurm.schedmd.com/slurm_ug_2012/MapRedSLURM.pdf http://slurm.schedmd.com/slurm_ug_2012/MapRedSLURM.pdf