Why do you say RSS feeds have no content to cache? There is the title, date and then, well, the content, either in the description element or content:encoded. And even in cases where it's just a title and a date, it takes time to open the connection and download the file.

The benefit is: once cached, do not have to connect to server and download the Web page. When there are 30 pages, this is an issue.

I'm already parallelizing the retrieval. I'm using LWP::Parallel after finding little additional speed benefit from either POE or HTTP::GHTTP with P::ForkManager.

Thanks anyway.