PROPELLER SUPER COMPUTING
MAKE MULTIPLE PROPELLER MACHINES FROM ONE
You can adapt this technique to make many machines given just one machine to work with. There's no need to tear down a machine for recycling or build a new machine with hardware. The method algorithm, which is all in software, is to first make a large parallel machine using Parallax Propeller chips and set the hardware according to the standard supercomputing wiring established by the Big Brain project. This is a parallel wiring format of chips and Cogs. Next, clone a large number of internal processors, i.e. 50,000 processors in a fifty chip machine and 100,000 processors in a one hundred chip machine. Now, configure the processors to make new machines. For example, each new machine is software wired, i.e. cube computing is possible, vast array machines are possible, new versions of T-Series Massive Transposition Machines, and linear ramp software accelerators are possible. Another example is the Power Cube, a machine inside a machine wired with node software.
How to wire a Propeller super computing machine with software is a very interesting vast topic and encompasses many areas. The subject can be addressed by examining reconfigurations of mass enumerated processors as a starting point. There are many techniques involving parallelism computing and distributed methodology. Look for more about this topic in the future.
