If you look for the way of making your expensive algorithm faster, you may consider multithreading and concurrency. You can divide your data into the parts and start threads to handle each part concurrently. This looks easy but sometimes there are some additional requisites that you need to obey doing that.

Solving some interesting problems in the project I’m currently working on I started to analyse following problem: you have a large data set (eg. a file) on which you need to apply some expensive transformation (eg. compression). The key requirement is a performance.