Beowulf Cluster finished

The Beowulf cluster we finished and the cluster worked beautifully, however it did not work with the cbc program which may or may not have been the cause of the bottleneck.  Our client gave us the .mps files that cbc (COIN-OR Branch and Cut) would use, however using mpich2 with cbc caused some sort of bottleneck where mpich2 couldn't parrallelize cbc and therefore the head node within the cluster solved the MILP contained within the .mps file.  This is what suspect, however, we are unable to prove this since we can't find a way to do that.  Perhaps in the future, a better program tracker can be developed for use with Linux.