In terms of performance you really need to minimize contention. If you build something that sends things round-robin to different spliterators you are going to have horrible performance, you really need to batch 10,000 or so lines and send those to different spliterators.
Also for anything that is pipelined there is ultimately going to be a bottleneck at the weakest point, which could be the spliterator factory.
Total failure, could not even get one core feed consistently because they where all fighting for resources.
Maybe could work for Non local or in memory data, but I rebuilt it with queues and a thread pool and it was easy way faster.