What are some best practices for optimizing memory management when working with large datasets?
I am tagging this topic both with php (because that is my language of choice, and the one I work with big data with) as well as c++ (because I know DaniWeb has a large low level c++ community that is well suited to being able to delve into this topic into depth, and because years ago when I focused on c++ myself, I was very focused on efficiency).