You know, I was not going to reply to your post, agreeing with it completely ... “There’s nothing novel here.”
Reading it a second time on my “ping list”, though, inspires me to note one thing that is really novel: that the cost of this Raspberry Pi unit dramatically lowers the cost of the effort to create a cluster like this. That is clearly novel, I believe.
As a veteran of many years in the supercomputing biz, here’s the thing: Cost is but one factor. What’s far more important: What is the task you are trying to work on, to pursue? What problem(s) are you trying to solve?
I’ll absolutely, positively guarantee you this.....my clients would laugh this off as a toy and then get back to real business. They deal with serious problems, serious science, and they have no truck with such bulls**t. DOE labs, major universities, govt. agencies, etc. These are the heavy hitters, and I know the technologies they need.
This cute little experiment ain’t it.