It's a long name I know. It's actually quite simple, and I suspect someone has done it before, but it would be nice to not only get a proof of concept, but also a working standard for programming a Parallel Computing Cluster. My idea is by using the GetCalc( function, Batlib/Hybrid BASIC etc. and two calculators, one can create two separate programs for each calculator, for the purpose of being able to use both calculators' processors to execute a specific task. For example, the easiest proof of concept I could think of was to make a parallel processing program that solved for prime numbers. The "mother" calculator would set up the system, then the other calculator (via link cable) would execute the necessary instructions that the mother calculator specifies. In this case, the mother calculator would use customary means to find the next probable prime number, then to check to see whether the number was prime, it would split the amount of checking it had to do with the other calcuator. Flow as follows:
…
Mother calculator finds that 101 is a probable prime number.
√101 is about 10 so:
Mother calculator splits that 10 in half to 5 so that:
The other calculator checks for divisibility between 5 of the numbers while:
The mother calculator checks for divisibility between the other 5 numbers.
Now in practice, checking the number 101 for primality only needs checking for divisibility by 2,3,5, and 7, but you get the idea. In the most optimized case, the mother would check for divisibility between 101 and 2-3, while the other would check for 5 and 7. While for a case as small as 101, there wouldn't be a very large increase in speed (if one at all) but when you end up needing to check hundreds of numbers to check for primality, being able to split the job would be very useful.
This sort of parallel processing is not just restricted to solving for prime numbers, but could end up being useful in a variety of different tasks. I plan on attempting to implement the scenario above. Tell me what you think!
