Parallel iterators on Fetcher do parallelism per item and let rayon take care of the rest. But it's probably faster to divide the work into core count chunks and then parallelize over those, since the amount of work per item is likely to be similar in most cases.
Parallel iterators on
Fetcherdo parallelism per item and letrayontake care of the rest. But it's probably faster to divide the work intocore countchunks and then parallelize over those, since the amount of work per item is likely to be similar in most cases.