Summary

In this chapter, we looked at three optimization techniques. The first technique involves finding the right algorithm and data structure. This has more impact on performance than any other single design or programming decision. Using the right algorithm can easily reduce runtimes from minutes to fractions of a second. Changing a poorly used sequence to a properly used mapping, for example, may reduce run time by a factor of 200.

We should generally optimize all of our recursions to be loops. This will be faster in Python and it won't be stopped by the call stack limit that Python imposes. There are many examples of how recursions are flattened into loops in other chapters, primarily, Chapter 6, Recursions and Reductions. Additionally, we may be able to improve performance in two other ways. First, we can apply memoization to cache results. For numeric calculations, this can have a large impact; for collections, the impact may be less. Secondly, replacing large materialized data...