Iterate over spins directly for multi-threading

This could have reduced memory usage by ensuring each thread only has to contain a single spin, but it did not. I am still going to keep it as it is cleaner code.