I have been planning to release an article with the exact same title, so naturally I hate the title. The article gives a quick insight on cache-aware algorithms

I have been planning to release an article with the exact same title, so naturally I hate the title. The article gives a quick insight on cache-aware algorithms, but misses to state that it's really about the data structures. Taking the example to the extreme, assuming you had huge, sparse matrices/vectors, how would you encode them? Or more generally, when does asymptotic complexity become a secondary criterion for the choice of a data structure? Otherwise, it's short but good, there can't be enough articles on that topic.