In this book, we show how to write very efficient Swift collection code. Throughout the book, we benchmark everything — with some surprising results. We implement custom data structures with value semantics and copy-on-write behavior such as sorted arrays, binary trees, red-black trees, and B-trees.
Even if you never implement your own collections, this book helps you reason about the performance of Swift code.
It's relatively short (roughly 170 pages on the iPad mini) and essentially walks you through a bunch of implementations of a sorted set in Swift. In the meantime, you learn about doing Copy on Write, Red-Black Trees, B-Trees and random chunks of Swift, algorithms and performance optimizations. It goes beyond nailing the asymptotic type and does some nice micro-optimizations that can help you learn about features of Swift that you might not be using. You won't end up with a lot of practical skills acquired from this book, but it is cool and enjoyable to read. If you don't know about Red-Black Trees and B-Trees, this is definitely a good place to learn a bit.
On the down side, it lacks a bit of completeness – it covers only inserting in trees, but not removing or the other operations (where it can get quite interesting). Still, it's great for such a short book.