OK, this sounds promissing. So what you're telling us is that clusterKNOPPIX will automatically distribute the work load over all the hardware on the cluster. You don't have to use any fancy compilers or write any fancy code for, say, massively parallel architectures or pipelined thruput, right? It just works?? This is VERY astounding!!!

Now, I'm installing a new network for my Computer Science students next year. I was thinking of using the thin-client tip elsewhere in the forum, but clustering has potential. OK, so what if I cluster 25 Pentium IV PC's, each running at 2Ghz with 512MB RAM and 100GB hdds. Will this run, say, one process as if it were a single 50Ghz machine with 12.5GB RAM and 2.5TB storage space????

TIA,