Waiting for a response from a servers is usually the limiting factor when it comes to scraping speed. This is especially the case when you want to scrape data from many different sites at once.
Ghostdriver makes it easy to manage a pool of phantomjs workers.
Ghostdriver launches a given phantomjs scraping script for a given range of urls in parallel/pooled or in series. A pool of child processes are spawned for each url for the given casperjs script.