I am interested in setting up a raspberry pi as the local end of a printing solution. In my testing the pi chokes on acting as a complete cups based print server. It seems a little underpowered for some of the ghostscipt processing and other filtering that occurs-- particularly on larger or complex documents the processing time can be 5 or more minutes.

My question is can the processing be largely done elsewhere and the prepared end product of the processing chain be fed to the pi for output on the connected printer?

So in this scenario any arbitrary document (html, pdf, text) is initially 'printed' on a relatively powerful machine but the output is stored in a file. This file is then grabbed by the pi and with all the heavy work out of the way easily printed using cups. I know files can be pushed through cups in raw mode but I am fuzzy on the pros and cons and the applicability in what I describe. I have tested this with pdftops creating a ps file then feeding that raw to cups and I think it works but it seems like there may be a cleaner solution.

This scenario would ideally work for any number of printer types.