This repo contains a few implementations of Conway's Game of Life. I mainly wrote these as a learning exercise in grid parallelization and Julia. They are probably not written particularly well but served their purpose. Some interesting results are:

• Parallelizing Python by starting every grid point on a new thread made it several orders of magnitude slower, if I added a sleep to the points' calculation function it was faster, so this could be useful for grids with very complex calculations.
• Splitting the grid sub grids with one sub grid for each CPU thread worked moderately well yielding a 6 times speed increase with 16 threads for an n=1280 grid (I would guess that this is because of the dispatch being slower than the actual grid computation)
• Rewriting in Julia was very easy and yielded a 20 times speed increase over the single threaded version
• Parallelization was then very easy again as you just have to add Threads.@threads before the for in a for loop and then run it with julia --threads n for n threads. This yielded a 3 times speed increase over the serial Julia version (again presumably because the grid is actually very simple)

An example grid:

The quotes speed increases are from runs with n=1280, 10 time steps and 16 threads for parallel versions. The results are in the table below: