Conceptual program, evaluated some efficiency in terms of CUDA Blocksize and number of threads


//Test notes:

Only allocating one block of threads results in small population size

Need to handle multiple blocks

Maximum size of block is 512 threads (more on other CUDA devices, but lets keep it safe)

Population size must be really big, i must try to fill as much memory as possible on the device

//Why is everything on one single file?

If people wanna try it out its easier this way, if it s one just one file it is more likely that they will try it, instead of importing and creating multiple files
Also, this is just a conceptual test, absolutely no intentions of realeasing it (I will start a new repo with the 'good' 'decent' and 'fast' version of this)