Comment by dnautics

I wonder if you could write parallelization routines that make this slightly faster with a few more atmels...