rayon implements the trait ParallelIterator for every type that implements its standard library equivalent Iterator, which we got to know in Chapter 2, Working with Collections; Access Collections as iterators. In fact, you can use all the knowledge from said recipe again here. The methods provided by the ParallelIterator trait are nearly the same as the ones provided by Iterator, so in virtually all cases where you notice an iterator operation taking too long and bottlenecking you, you can simply replace .iter() with .par_iter()[10]. Similarly, for moving iterators, you can use .into_par_iter() instead of .into_iter().

rayon handles all the tedious work for you, as it automatically distributes the work evenly between all of your available cores. Just keep in mind that despite this magic, you're still dealing with parallelism here, so you have no guarantees about the order in which the items in your iterator are going to be handled, as evidenced by line [10], which will print in a different order each time you execute the program:

words.par_iter().for_each(|val| println!("{}", val));